Discrete wavelength selection for the optical readout of a metamaterial biosensing system for glucose concentration estimation via a support vector regression model.

PubMed

Teutsch, T; Mesch, M; Giessen, H; Tarin, C

2015-01-01

In this contribution, a method to select discrete wavelengths that allow an accurate estimation of the glucose concentration in a biosensing system based on metamaterials is presented. The sensing concept is adapted to the particular application of ophthalmic glucose sensing by covering the metamaterial with a glucose-sensitive hydrogel and the sensor readout is performed optically. Due to the fact that in a mobile context a spectrometer is not suitable, few discrete wavelengths must be selected to estimate the glucose concentration. The developed selection methods are based on nonlinear support vector regression (SVR) models. Two selection methods are compared and it is shown that wavelengths selected by a sequential forward feature selection algorithm achieves an estimation improvement. The presented method can be easily applied to different metamaterial layouts and hydrogel configurations.

PREDICTION OF MALIGNANT BREAST LESIONS FROM MRI FEATURES: A COMPARISON OF ARTIFICIAL NEURAL NETWORK AND LOGISTIC REGRESSION TECHNIQUES

PubMed Central

McLaren, Christine E.; Chen, Wen-Pin; Nie, Ke; Su, Min-Ying

2009-01-01

Rationale and Objectives Dynamic contrast enhanced MRI (DCE-MRI) is a clinical imaging modality for detection and diagnosis of breast lesions. Analytical methods were compared for diagnostic feature selection and performance of lesion classification to differentiate between malignant and benign lesions in patients. Materials and Methods The study included 43 malignant and 28 benign histologically-proven lesions. Eight morphological parameters, ten gray level co-occurrence matrices (GLCM) texture features, and fourteen Laws’ texture features were obtained using automated lesion segmentation and quantitative feature extraction. Artificial neural network (ANN) and logistic regression analysis were compared for selection of the best predictors of malignant lesions among the normalized features. Results Using ANN, the final four selected features were compactness, energy, homogeneity, and Law_LS, with area under the receiver operating characteristic curve (AUC) = 0.82, and accuracy = 0.76. The diagnostic performance of these 4-features computed on the basis of logistic regression yielded AUC = 0.80 (95% CI, 0.688 to 0.905), similar to that of ANN. The analysis also shows that the odds of a malignant lesion decreased by 48% (95% CI, 25% to 92%) for every increase of 1 SD in the Law_LS feature, adjusted for differences in compactness, energy, and homogeneity. Using logistic regression with z-score transformation, a model comprised of compactness, NRL entropy, and gray level sum average was selected, and it had the highest overall accuracy of 0.75 among all models, with AUC = 0.77 (95% CI, 0.660 to 0.880). When logistic modeling of transformations using the Box-Cox method was performed, the most parsimonious model with predictors, compactness and Law_LS, had an AUC of 0.79 (95% CI, 0.672 to 0.898). Conclusion The diagnostic performance of models selected by ANN and logistic regression was similar. The analytic methods were found to be roughly equivalent in terms of predictive ability when a small number of variables were chosen. The robust ANN methodology utilizes a sophisticated non-linear model, while logistic regression analysis provides insightful information to enhance interpretation of the model features. PMID:19409817

A Survey of UML Based Regression Testing

NASA Astrophysics Data System (ADS)

Fahad, Muhammad; Nadeem, Aamer

Regression testing is the process of ensuring software quality by analyzing whether changed parts behave as intended, and unchanged parts are not affected by the modifications. Since it is a costly process, a lot of techniques are proposed in the research literature that suggest testers how to build regression test suite from existing test suite with minimum cost. In this paper, we discuss the advantages and drawbacks of using UML diagrams for regression testing and analyze that UML model helps in identifying changes for regression test selection effectively. We survey the existing UML based regression testing techniques and provide an analysis matrix to give a quick insight into prominent features of the literature work. We discuss the open research issues like managing and reducing the size of regression test suite, prioritization of the test cases that would be helpful during strict schedule and resources that remain to be addressed for UML based regression testing.

Perceptual quality estimation of H.264/AVC videos using reduced-reference and no-reference models

NASA Astrophysics Data System (ADS)

Shahid, Muhammad; Pandremmenou, Katerina; Kondi, Lisimachos P.; Rossholm, Andreas; Lövström, Benny

2016-09-01

Reduced-reference (RR) and no-reference (NR) models for video quality estimation, using features that account for the impact of coding artifacts, spatio-temporal complexity, and packet losses, are proposed. The purpose of this study is to analyze a number of potentially quality-relevant features in order to select the most suitable set of features for building the desired models. The proposed sets of features have not been used in the literature and some of the features are used for the first time in this study. The features are employed by the least absolute shrinkage and selection operator (LASSO), which selects only the most influential of them toward perceptual quality. For comparison, we apply feature selection in the complete feature sets and ridge regression on the reduced sets. The models are validated using a database of H.264/AVC encoded videos that were subjectively assessed for quality in an ITU-T compliant laboratory. We infer that just two features selected by RR LASSO and two bitstream-based features selected by NR LASSO are able to estimate perceptual quality with high accuracy, higher than that of ridge, which uses more features. The comparisons with competing works and two full-reference metrics also verify the superiority of our models.

Optimizing methods for linking cinematic features to fMRI data.

PubMed

Kauttonen, Janne; Hlushchuk, Yevhen; Tikka, Pia

2015-04-15

One of the challenges of naturalistic neurosciences using movie-viewing experiments is how to interpret observed brain activations in relation to the multiplicity of time-locked stimulus features. As previous studies have shown less inter-subject synchronization across viewers of random video footage than story-driven films, new methods need to be developed for analysis of less story-driven contents. To optimize the linkage between our fMRI data collected during viewing of a deliberately non-narrative silent film 'At Land' by Maya Deren (1944) and its annotated content, we combined the method of elastic-net regularization with the model-driven linear regression and the well-established data-driven independent component analysis (ICA) and inter-subject correlation (ISC) methods. In the linear regression analysis, both IC and region-of-interest (ROI) time-series were fitted with time-series of a total of 36 binary-valued and one real-valued tactile annotation of film features. The elastic-net regularization and cross-validation were applied in the ordinary least-squares linear regression in order to avoid over-fitting due to the multicollinearity of regressors, the results were compared against both the partial least-squares (PLS) regression and the un-regularized full-model regression. Non-parametric permutation testing scheme was applied to evaluate the statistical significance of regression. We found statistically significant correlation between the annotation model and 9 ICs out of 40 ICs. Regression analysis was also repeated for a large set of cubic ROIs covering the grey matter. Both IC- and ROI-based regression analyses revealed activations in parietal and occipital regions, with additional smaller clusters in the frontal lobe. Furthermore, we found elastic-net based regression more sensitive than PLS and un-regularized regression since it detected a larger number of significant ICs and ROIs. Along with the ISC ranking methods, our regression analysis proved a feasible method for ordering the ICs based on their functional relevance to the annotated cinematic features. The novelty of our method is - in comparison to the hypothesis-driven manual pre-selection and observation of some individual regressors biased by choice - in applying data-driven approach to all content features simultaneously. We found especially the combination of regularized regression and ICA useful when analyzing fMRI data obtained using non-narrative movie stimulus with a large set of complex and correlated features. Copyright © 2015. Published by Elsevier Inc.

Variable selection in near-infrared spectroscopy: benchmarking of feature selection methods on biodiesel data.

PubMed

Balabin, Roman M; Smirnov, Sergey V

2011-04-29

During the past several years, near-infrared (near-IR/NIR) spectroscopy has increasingly been adopted as an analytical tool in various fields from petroleum to biomedical sectors. The NIR spectrum (above 4000 cm(-1)) of a sample is typically measured by modern instruments at a few hundred of wavelengths. Recently, considerable effort has been directed towards developing procedures to identify variables (wavelengths) that contribute useful information. Variable selection (VS) or feature selection, also called frequency selection or wavelength selection, is a critical step in data analysis for vibrational spectroscopy (infrared, Raman, or NIRS). In this paper, we compare the performance of 16 different feature selection methods for the prediction of properties of biodiesel fuel, including density, viscosity, methanol content, and water concentration. The feature selection algorithms tested include stepwise multiple linear regression (MLR-step), interval partial least squares regression (iPLS), backward iPLS (BiPLS), forward iPLS (FiPLS), moving window partial least squares regression (MWPLS), (modified) changeable size moving window partial least squares (CSMWPLS/MCSMWPLSR), searching combination moving window partial least squares (SCMWPLS), successive projections algorithm (SPA), uninformative variable elimination (UVE, including UVE-SPA), simulated annealing (SA), back-propagation artificial neural networks (BP-ANN), Kohonen artificial neural network (K-ANN), and genetic algorithms (GAs, including GA-iPLS). Two linear techniques for calibration model building, namely multiple linear regression (MLR) and partial least squares regression/projection to latent structures (PLS/PLSR), are used for the evaluation of biofuel properties. A comparison with a non-linear calibration model, artificial neural networks (ANN-MLP), is also provided. Discussion of gasoline, ethanol-gasoline (bioethanol), and diesel fuel data is presented. The results of other spectroscopic techniques application, such as Raman, ultraviolet-visible (UV-vis), or nuclear magnetic resonance (NMR) spectroscopies, can be greatly improved by an appropriate feature selection choice. Copyright © 2011 Elsevier B.V. All rights reserved.

Integrative approach for inference of gene regulatory networks using lasso-based random featuring and application to psychiatric disorders.

PubMed

Kim, Dongchul; Kang, Mingon; Biswas, Ashis; Liu, Chunyu; Gao, Jean

2016-08-10

Inferring gene regulatory networks is one of the most interesting research areas in the systems biology. Many inference methods have been developed by using a variety of computational models and approaches. However, there are two issues to solve. First, depending on the structural or computational model of inference method, the results tend to be inconsistent due to innately different advantages and limitations of the methods. Therefore the combination of dissimilar approaches is demanded as an alternative way in order to overcome the limitations of standalone methods through complementary integration. Second, sparse linear regression that is penalized by the regularization parameter (lasso) and bootstrapping-based sparse linear regression methods were suggested in state of the art methods for network inference but they are not effective for a small sample size data and also a true regulator could be missed if the target gene is strongly affected by an indirect regulator with high correlation or another true regulator. We present two novel network inference methods based on the integration of three different criteria, (i) z-score to measure the variation of gene expression from knockout data, (ii) mutual information for the dependency between two genes, and (iii) linear regression-based feature selection. Based on these criterion, we propose a lasso-based random feature selection algorithm (LARF) to achieve better performance overcoming the limitations of bootstrapping as mentioned above. In this work, there are three main contributions. First, our z score-based method to measure gene expression variations from knockout data is more effective than similar criteria of related works. Second, we confirmed that the true regulator selection can be effectively improved by LARF. Lastly, we verified that an integrative approach can clearly outperform a single method when two different methods are effectively jointed. In the experiments, our methods were validated by outperforming the state of the art methods on DREAM challenge data, and then LARF was applied to inferences of gene regulatory network associated with psychiatric disorders.

Identifying key radiogenomic associations between DCE-MRI and micro-RNA expressions for breast cancer

NASA Astrophysics Data System (ADS)

Samala, Ravi K.; Chan, Heang-Ping; Hadjiiski, Lubomir; Helvie, Mark A.; Kim, Renaid

2017-03-01

Understanding the key radiogenomic associations for breast cancer between DCE-MRI and micro-RNA expressions is the foundation for the discovery of radiomic features as biomarkers for assessing tumor progression and prognosis. We conducted a study to analyze the radiogenomic associations for breast cancer using the TCGA-TCIA data set. The core idea that tumor etiology is a function of the behavior of miRNAs is used to build the regression models. The associations based on regression are analyzed for three study outcomes: diagnosis, prognosis, and treatment. The diagnosis group consists of miRNAs associated with clinicopathologic features of breast cancer and significant aberration of expression in breast cancer patients. The prognosis group consists of miRNAs which are closely associated with tumor suppression and regulation of cell proliferation and differentiation. The treatment group consists of miRNAs that contribute significantly to the regulation of metastasis thereby having the potential to be part of therapeutic mechanisms. As a first step, important miRNA expressions were identified and their ability to classify the clinical phenotypes based on the study outcomes was evaluated using the area under the ROC curve (AUC) as a figure-of-merit. The key mapping between the selected miRNAs and radiomic features were determined using least absolute shrinkage and selection operator (LASSO) regression analysis within a two-loop leave-one-out cross-validation strategy. These key associations indicated a number of radiomic features from DCE-MRI to be potential biomarkers for the three study outcomes.

SU-E-J-256: Predicting Metastasis-Free Survival of Rectal Cancer Patients Treated with Neoadjuvant Chemo-Radiotherapy by Data-Mining of CT Texture Features of Primary Lesions

DOE Office of Scientific and Technical Information (OSTI.GOV)

Zhong, H; Wang, J; Shen, L

Purpose: The purpose of this study is to investigate the relationship between computed tomographic (CT) texture features of primary lesions and metastasis-free survival for rectal cancer patients; and to develop a datamining prediction model using texture features. Methods: A total of 220 rectal cancer patients treated with neoadjuvant chemo-radiotherapy (CRT) were enrolled in this study. All patients underwent CT scans before CRT. The primary lesions on the CT images were delineated by two experienced oncologists. The CT images were filtered by Laplacian of Gaussian (LoG) filters with different filter values (1.0–2.5: from fine to coarse). Both filtered and unfiltered imagesmore » were analyzed using Gray-level Co-occurrence Matrix (GLCM) texture analysis with different directions (transversal, sagittal, and coronal). Totally, 270 texture features with different species, directions and filter values were extracted. Texture features were examined with Student’s t-test for selecting predictive features. Principal Component Analysis (PCA) was performed upon the selected features to reduce the feature collinearity. Artificial neural network (ANN) and logistic regression were applied to establish metastasis prediction models. Results: Forty-six of 220 patients developed metastasis with a follow-up time of more than 2 years. Sixtyseven texture features were significantly different in t-test (p<0.05) between patients with and without metastasis, and 12 of them were extremely significant (p<0.001). The Area-under-the-curve (AUC) of ANN was 0.72, and the concordance index (CI) of logistic regression was 0.71. The predictability of ANN was slightly better than logistic regression. Conclusion: CT texture features of primary lesions are related to metastasisfree survival of rectal cancer patients. Both ANN and logistic regression based models can be developed for prediction.« less

Characterizing mammographic images by using generic texture features

PubMed Central

2012-01-01

Introduction Although mammographic density is an established risk factor for breast cancer, its use is limited in clinical practice because of a lack of automated and standardized measurement methods. The aims of this study were to evaluate a variety of automated texture features in mammograms as risk factors for breast cancer and to compare them with the percentage mammographic density (PMD) by using a case-control study design. Methods A case-control study including 864 cases and 418 controls was analyzed automatically. Four hundred seventy features were explored as possible risk factors for breast cancer. These included statistical features, moment-based features, spectral-energy features, and form-based features. An elaborate variable selection process using logistic regression analyses was performed to identify those features that were associated with case-control status. In addition, PMD was assessed and included in the regression model. Results Of the 470 image-analysis features explored, 46 remained in the final logistic regression model. An area under the curve of 0.79, with an odds ratio per standard deviation change of 2.88 (95% CI, 2.28 to 3.65), was obtained with validation data. Adding the PMD did not improve the final model. Conclusions Using texture features to predict the risk of breast cancer appears feasible. PMD did not show any additional value in this study. With regard to the features assessed, most of the analysis tools appeared to reflect mammographic density, although some features did not correlate with PMD. It remains to be investigated in larger case-control studies whether these features can contribute to increased prediction accuracy. PMID:22490545

Automated system for characterization and classification of malaria-infected stages using light microscopic images of thin blood smears.

PubMed

Das, D K; Maiti, A K; Chakraborty, C

2015-03-01

In this paper, we propose a comprehensive image characterization cum classification framework for malaria-infected stage detection using microscopic images of thin blood smears. The methodology mainly includes microscopic imaging of Leishman stained blood slides, noise reduction and illumination correction, erythrocyte segmentation, feature selection followed by machine classification. Amongst three-image segmentation algorithms (namely, rule-based, Chan-Vese-based and marker-controlled watershed methods), marker-controlled watershed technique provides better boundary detection of erythrocytes specially in overlapping situations. Microscopic features at intensity, texture and morphology levels are extracted to discriminate infected and noninfected erythrocytes. In order to achieve subgroup of potential features, feature selection techniques, namely, F-statistic and information gain criteria are considered here for ranking. Finally, five different classifiers, namely, Naive Bayes, multilayer perceptron neural network, logistic regression, classification and regression tree (CART), RBF neural network have been trained and tested by 888 erythrocytes (infected and noninfected) for each features' subset. Performance evaluation of the proposed methodology shows that multilayer perceptron network provides higher accuracy for malaria-infected erythrocytes recognition and infected stage classification. Results show that top 90 features ranked by F-statistic (specificity: 98.64%, sensitivity: 100%, PPV: 99.73% and overall accuracy: 96.84%) and top 60 features ranked by information gain provides better results (specificity: 97.29%, sensitivity: 100%, PPV: 99.46% and overall accuracy: 96.73%) for malaria-infected stage classification. © 2014 The Authors Journal of Microscopy © 2014 Royal Microscopical Society.

Influences on choice of surgery as a career: a study of consecutive cohorts in a medical school.

PubMed

Sobral, Dejano T

2006-06-01

To examine the differential impact of person-based and programme-related features on graduates' dichotomous choice between surgical or non-surgical field specialties for first-year residency. A 10-year cohort study was conducted, following 578 students (55.4% male) who graduated from a university medical school during 1994-2003. Data were collected as follows: at the beginning of medical studies, on career preference and learning frame; during medical studies, on academic achievement, cross-year peer tutoring and selective clinical traineeship, and at graduation, on the first-year residency selected. Contingency and logistic regression analyses were performed, with graduates grouped by the dichotomous choice of surgery or not. Overall, 23% of graduates selected a first-year residency in surgery. Seven time-steady features related to this choice: male sex, high self-confidence, option of surgery at admission, active learning style, preference for surgery after Year 1, peer tutoring on clinical surgery, and selective training in clinical surgery. Logistic regression analysis, including all features, predicted 87.1% of the graduates' choices. Male sex, updated preference, peer tutoring and selective training were the most significant predictors in the pathway to choice. The relative roles of person-based and programme-related factors in the choice process are discussed. The findings suggest that for most students the choice of surgery derives from a temporal summation of influences that encompass entry and post-entry factors blended in variable patterns. It is likely that sex-unbiased peer tutoring and selective training supported the students' search process for personal compatibility with specialty-related domains of content and process.

A P2P Botnet detection scheme based on decision tree and adaptive multilayer neural networks.

PubMed

Alauthaman, Mohammad; Aslam, Nauman; Zhang, Li; Alasem, Rafe; Hossain, M A

2018-01-01

In recent years, Botnets have been adopted as a popular method to carry and spread many malicious codes on the Internet. These malicious codes pave the way to execute many fraudulent activities including spam mail, distributed denial-of-service attacks and click fraud. While many Botnets are set up using centralized communication architecture, the peer-to-peer (P2P) Botnets can adopt a decentralized architecture using an overlay network for exchanging command and control data making their detection even more difficult. This work presents a method of P2P Bot detection based on an adaptive multilayer feed-forward neural network in cooperation with decision trees. A classification and regression tree is applied as a feature selection technique to select relevant features. With these features, a multilayer feed-forward neural network training model is created using a resilient back-propagation learning algorithm. A comparison of feature set selection based on the decision tree, principal component analysis and the ReliefF algorithm indicated that the neural network model with features selection based on decision tree has a better identification accuracy along with lower rates of false positives. The usefulness of the proposed approach is demonstrated by conducting experiments on real network traffic datasets. In these experiments, an average detection rate of 99.08 % with false positive rate of 0.75 % was observed.

Appropriateness guidelines and predictive rules to select patients for upper endoscopy: a nationwide multicenter study.

PubMed

Buri, Luigi; Hassan, Cesare; Bersani, Gianluca; Anti, Marcello; Bianco, Maria Antonietta; Cipolletta, Livio; Di Giulio, Emilio; Di Matteo, Giovanni; Familiari, Luigi; Ficano, Leonardo; Loriga, Pietro; Morini, Sergio; Pietropaolo, Vincenzo; Zambelli, Alessandro; Grossi, Enzo; Intraligi, Marco; Buscema, Massimo

2010-06-01

Selecting patients appropriately for upper endoscopy (EGD) is crucial for efficient use of endoscopy. The objective of this study was to compare different clinical strategies and statistical methods to select patients for EGD, namely appropriateness guidelines, age and/or alarm features, and multivariate and artificial neural network (ANN) models. A nationwide, multicenter, prospective study was undertaken in which consecutive patients referred for EGD during a 1-month period were enrolled. Before EGD, the endoscopist assessed referral appropriateness according to the American Society for Gastrointestinal Endoscopy (ASGE) guidelines, also collecting clinical and demographic variables. Outcomes of the study were detection of relevant findings and new diagnosis of malignancy at EGD. The accuracy of the following clinical strategies and predictive rules was compared: (i) ASGE appropriateness guidelines (indicated vs. not indicated), (ii) simplified rule (>or=45 years or alarm features vs. <45 years without alarm features), (iii) logistic regression model, and (iv) ANN models. A total of 8,252 patients were enrolled in 57 centers. Overall, 3,803 (46%) relevant findings and 132 (1.6%) new malignancies were detected. Sensitivity, specificity, and area under the receiver-operating characteristic curve (AUC) of the simplified rule were similar to that of the ASGE guidelines for both relevant findings (82%/26%/0.55 vs. 88%/27%/0.52) and cancer (97%/22%/0.58 vs. 98%/20%/0.58). Both logistic regression and ANN models seemed to be substantially more accurate in predicting new cases of malignancy, with an AUC of 0.82 and 0.87, respectively. A simple predictive rule based on age and alarm features is similarly effective to the more complex ASGE guidelines in selecting patients for EGD. Regression and ANN models may be useful in identifying a relatively small subgroup of patients at higher risk of cancer.

Oral cancer prognosis based on clinicopathologic and genomic markers using a hybrid of feature selection and machine learning methods

PubMed Central

2013-01-01

Background Machine learning techniques are becoming useful as an alternative approach to conventional medical diagnosis or prognosis as they are good for handling noisy and incomplete data, and significant results can be attained despite a small sample size. Traditionally, clinicians make prognostic decisions based on clinicopathologic markers. However, it is not easy for the most skilful clinician to come out with an accurate prognosis by using these markers alone. Thus, there is a need to use genomic markers to improve the accuracy of prognosis. The main aim of this research is to apply a hybrid of feature selection and machine learning methods in oral cancer prognosis based on the parameters of the correlation of clinicopathologic and genomic markers. Results In the first stage of this research, five feature selection methods have been proposed and experimented on the oral cancer prognosis dataset. In the second stage, the model with the features selected from each feature selection methods are tested on the proposed classifiers. Four types of classifiers are chosen; these are namely, ANFIS, artificial neural network, support vector machine and logistic regression. A k-fold cross-validation is implemented on all types of classifiers due to the small sample size. The hybrid model of ReliefF-GA-ANFIS with 3-input features of drink, invasion and p63 achieved the best accuracy (accuracy = 93.81%; AUC = 0.90) for the oral cancer prognosis. Conclusions The results revealed that the prognosis is superior with the presence of both clinicopathologic and genomic markers. The selected features can be investigated further to validate the potential of becoming as significant prognostic signature in the oral cancer studies. PMID:23725313

Compensatory selection for roads over natural linear features by wolves in northern Ontario: Implications for caribou conservation

PubMed Central

Patterson, Brent R.; Anderson, Morgan L.; Rodgers, Arthur R.; Vander Vennen, Lucas M.; Fryxell, John M.

2017-01-01

Woodland caribou (Rangifer tarandus caribou) in Ontario are a threatened species that have experienced a substantial retraction of their historic range. Part of their decline has been attributed to increasing densities of anthropogenic linear features such as trails, roads, railways, and hydro lines. These features have been shown to increase the search efficiency and kill rate of wolves. However, it is unclear whether selection for anthropogenic linear features is additive or compensatory to selection for natural (water) linear features which may also be used for travel. We studied the selection of water and anthropogenic linear features by 52 resident wolves (Canis lupus x lycaon) over four years across three study areas in northern Ontario that varied in degrees of forestry activity and human disturbance. We used Euclidean distance-based resource selection functions (mixed-effects logistic regression) at the seasonal range scale with random coefficients for distance to water linear features, primary/secondary roads/railways, and hydro lines, and tertiary roads to estimate the strength of selection for each linear feature and for several habitat types, while accounting for availability of each feature. Next, we investigated the trade-off between selection for anthropogenic and water linear features. Wolves selected both anthropogenic and water linear features; selection for anthropogenic features was stronger than for water during the rendezvous season. Selection for anthropogenic linear features increased with increasing density of these features on the landscape, while selection for natural linear features declined, indicating compensatory selection of anthropogenic linear features. These results have implications for woodland caribou conservation. Prey encounter rates between wolves and caribou seem to be strongly influenced by increasing linear feature densities. This behavioral mechanism–a compensatory functional response to anthropogenic linear feature density resulting in decreased use of natural travel corridors–has negative consequences for the viability of woodland caribou. PMID:29117234

Compensatory selection for roads over natural linear features by wolves in northern Ontario: Implications for caribou conservation.

PubMed

Newton, Erica J; Patterson, Brent R; Anderson, Morgan L; Rodgers, Arthur R; Vander Vennen, Lucas M; Fryxell, John M

2017-01-01

Woodland caribou (Rangifer tarandus caribou) in Ontario are a threatened species that have experienced a substantial retraction of their historic range. Part of their decline has been attributed to increasing densities of anthropogenic linear features such as trails, roads, railways, and hydro lines. These features have been shown to increase the search efficiency and kill rate of wolves. However, it is unclear whether selection for anthropogenic linear features is additive or compensatory to selection for natural (water) linear features which may also be used for travel. We studied the selection of water and anthropogenic linear features by 52 resident wolves (Canis lupus x lycaon) over four years across three study areas in northern Ontario that varied in degrees of forestry activity and human disturbance. We used Euclidean distance-based resource selection functions (mixed-effects logistic regression) at the seasonal range scale with random coefficients for distance to water linear features, primary/secondary roads/railways, and hydro lines, and tertiary roads to estimate the strength of selection for each linear feature and for several habitat types, while accounting for availability of each feature. Next, we investigated the trade-off between selection for anthropogenic and water linear features. Wolves selected both anthropogenic and water linear features; selection for anthropogenic features was stronger than for water during the rendezvous season. Selection for anthropogenic linear features increased with increasing density of these features on the landscape, while selection for natural linear features declined, indicating compensatory selection of anthropogenic linear features. These results have implications for woodland caribou conservation. Prey encounter rates between wolves and caribou seem to be strongly influenced by increasing linear feature densities. This behavioral mechanism-a compensatory functional response to anthropogenic linear feature density resulting in decreased use of natural travel corridors-has negative consequences for the viability of woodland caribou.

Tehran Air Pollutants Prediction Based on Random Forest Feature Selection Method

NASA Astrophysics Data System (ADS)

Shamsoddini, A.; Aboodi, M. R.; Karami, J.

2017-09-01

Air pollution as one of the most serious forms of environmental pollutions poses huge threat to human life. Air pollution leads to environmental instability, and has harmful and undesirable effects on the environment. Modern prediction methods of the pollutant concentration are able to improve decision making and provide appropriate solutions. This study examines the performance of the Random Forest feature selection in combination with multiple-linear regression and Multilayer Perceptron Artificial Neural Networks methods, in order to achieve an efficient model to estimate carbon monoxide and nitrogen dioxide, sulfur dioxide and PM2.5 contents in the air. The results indicated that Artificial Neural Networks fed by the attributes selected by Random Forest feature selection method performed more accurate than other models for the modeling of all pollutants. The estimation accuracy of sulfur dioxide emissions was lower than the other air contaminants whereas the nitrogen dioxide was predicted more accurate than the other pollutants.

Prediction models for solitary pulmonary nodules based on curvelet textural features and clinical parameters.

PubMed

Wang, Jing-Jing; Wu, Hai-Feng; Sun, Tao; Li, Xia; Wang, Wei; Tao, Li-Xin; Huo, Da; Lv, Ping-Xin; He, Wen; Guo, Xiu-Hua

2013-01-01

Lung cancer, one of the leading causes of cancer-related deaths, usually appears as solitary pulmonary nodules (SPNs) which are hard to diagnose using the naked eye. In this paper, curvelet-based textural features and clinical parameters are used with three prediction models [a multilevel model, a least absolute shrinkage and selection operator (LASSO) regression method, and a support vector machine (SVM)] to improve the diagnosis of benign and malignant SPNs. Dimensionality reduction of the original curvelet-based textural features was achieved using principal component analysis. In addition, non-conditional logistical regression was used to find clinical predictors among demographic parameters and morphological features. The results showed that, combined with 11 clinical predictors, the accuracy rates using 12 principal components were higher than those using the original curvelet-based textural features. To evaluate the models, 10-fold cross validation and back substitution were applied. The results obtained, respectively, were 0.8549 and 0.9221 for the LASSO method, 0.9443 and 0.9831 for SVM, and 0.8722 and 0.9722 for the multilevel model. All in all, it was found that using curvelet-based textural features after dimensionality reduction and using clinical predictors, the highest accuracy rate was achieved with SVM. The method may be used as an auxiliary tool to differentiate between benign and malignant SPNs in CT images.

Feature selection in feature network models: finding predictive subsets of features with the Positive Lasso.

PubMed

Frank, Laurence E; Heiser, Willem J

2008-05-01

A set of features is the basis for the network representation of proximity data achieved by feature network models (FNMs). Features are binary variables that characterize the objects in an experiment, with some measure of proximity as response variable. Sometimes features are provided by theory and play an important role in the construction of the experimental conditions. In some research settings, the features are not known a priori. This paper shows how to generate features in this situation and how to select an adequate subset of features that takes into account a good compromise between model fit and model complexity, using a new version of least angle regression that restricts coefficients to be non-negative, called the Positive Lasso. It will be shown that features can be generated efficiently with Gray codes that are naturally linked to the FNMs. The model selection strategy makes use of the fact that FNM can be considered as univariate multiple regression model. A simulation study shows that the proposed strategy leads to satisfactory results if the number of objects is less than or equal to 22. If the number of objects is larger than 22, the number of features selected by our method exceeds the true number of features in some conditions.

Identification of Alfalfa Leaf Diseases Using Image Recognition Technology

PubMed Central

Qin, Feng; Liu, Dongxia; Sun, Bingda; Ruan, Liu; Ma, Zhanhong; Wang, Haiguang

2016-01-01

Common leaf spot (caused by Pseudopeziza medicaginis), rust (caused by Uromyces striatus), Leptosphaerulina leaf spot (caused by Leptosphaerulina briosiana) and Cercospora leaf spot (caused by Cercospora medicaginis) are the four common types of alfalfa leaf diseases. Timely and accurate diagnoses of these diseases are critical for disease management, alfalfa quality control and the healthy development of the alfalfa industry. In this study, the identification and diagnosis of the four types of alfalfa leaf diseases were investigated using pattern recognition algorithms based on image-processing technology. A sub-image with one or multiple typical lesions was obtained by artificial cutting from each acquired digital disease image. Then the sub-images were segmented using twelve lesion segmentation methods integrated with clustering algorithms (including K_means clustering, fuzzy C-means clustering and K_median clustering) and supervised classification algorithms (including logistic regression analysis, Naive Bayes algorithm, classification and regression tree, and linear discriminant analysis). After a comprehensive comparison, the segmentation method integrating the K_median clustering algorithm and linear discriminant analysis was chosen to obtain lesion images. After the lesion segmentation using this method, a total of 129 texture, color and shape features were extracted from the lesion images. Based on the features selected using three methods (ReliefF, 1R and correlation-based feature selection), disease recognition models were built using three supervised learning methods, including the random forest, support vector machine (SVM) and K-nearest neighbor methods. A comparison of the recognition results of the models was conducted. The results showed that when the ReliefF method was used for feature selection, the SVM model built with the most important 45 features (selected from a total of 129 features) was the optimal model. For this SVM model, the recognition accuracies of the training set and the testing set were 97.64% and 94.74%, respectively. Semi-supervised models for disease recognition were built based on the 45 effective features that were used for building the optimal SVM model. For the optimal semi-supervised models built with three ratios of labeled to unlabeled samples in the training set, the recognition accuracies of the training set and the testing set were both approximately 80%. The results indicated that image recognition of the four alfalfa leaf diseases can be implemented with high accuracy. This study provides a feasible solution for lesion image segmentation and image recognition of alfalfa leaf disease. PMID:27977767

Identification of Alfalfa Leaf Diseases Using Image Recognition Technology.

PubMed

Qin, Feng; Liu, Dongxia; Sun, Bingda; Ruan, Liu; Ma, Zhanhong; Wang, Haiguang

2016-01-01

Common leaf spot (caused by Pseudopeziza medicaginis), rust (caused by Uromyces striatus), Leptosphaerulina leaf spot (caused by Leptosphaerulina briosiana) and Cercospora leaf spot (caused by Cercospora medicaginis) are the four common types of alfalfa leaf diseases. Timely and accurate diagnoses of these diseases are critical for disease management, alfalfa quality control and the healthy development of the alfalfa industry. In this study, the identification and diagnosis of the four types of alfalfa leaf diseases were investigated using pattern recognition algorithms based on image-processing technology. A sub-image with one or multiple typical lesions was obtained by artificial cutting from each acquired digital disease image. Then the sub-images were segmented using twelve lesion segmentation methods integrated with clustering algorithms (including K_means clustering, fuzzy C-means clustering and K_median clustering) and supervised classification algorithms (including logistic regression analysis, Naive Bayes algorithm, classification and regression tree, and linear discriminant analysis). After a comprehensive comparison, the segmentation method integrating the K_median clustering algorithm and linear discriminant analysis was chosen to obtain lesion images. After the lesion segmentation using this method, a total of 129 texture, color and shape features were extracted from the lesion images. Based on the features selected using three methods (ReliefF, 1R and correlation-based feature selection), disease recognition models were built using three supervised learning methods, including the random forest, support vector machine (SVM) and K-nearest neighbor methods. A comparison of the recognition results of the models was conducted. The results showed that when the ReliefF method was used for feature selection, the SVM model built with the most important 45 features (selected from a total of 129 features) was the optimal model. For this SVM model, the recognition accuracies of the training set and the testing set were 97.64% and 94.74%, respectively. Semi-supervised models for disease recognition were built based on the 45 effective features that were used for building the optimal SVM model. For the optimal semi-supervised models built with three ratios of labeled to unlabeled samples in the training set, the recognition accuracies of the training set and the testing set were both approximately 80%. The results indicated that image recognition of the four alfalfa leaf diseases can be implemented with high accuracy. This study provides a feasible solution for lesion image segmentation and image recognition of alfalfa leaf disease.

Discriminating Induced-Microearthquakes Using New Seismic Features

NASA Astrophysics Data System (ADS)

Mousavi, S. M.; Horton, S.

2016-12-01

We studied characteristics of induced-microearthquakes on the basis of the waveforms recorded on a limited number of surface receivers using machine-learning techniques. Forty features in the time, frequency, and time-frequency domains were measured on each waveform, and several techniques such as correlation-based feature selection, Artificial Neural Networks (ANNs), Logistic Regression (LR) and X-mean were used as research tools to explore the relationship between these seismic features and source parameters. The results show that spectral features have the highest correlation to source depth. Two new measurements developed as seismic features for this study, spectral centroids and 2D cross-correlations in the time-frequency domain, performed better than the common seismic measurements. These features can be used by machine learning techniques for efficient automatic classification of low energy signals recorded at one or more seismic stations. We applied the technique to 440 microearthquakes-1.7Reference: Mousavi, S.M., S.P. Horton, C. A. Langston, B. Samei, (2016) Seismic features and automatic discrimination of deep and shallow induced-microearthquakes using neural network and logistic regression, Geophys. J. Int. doi: 10.1093/gji/ggw258.

Complex Environmental Data Modelling Using Adaptive General Regression Neural Networks

NASA Astrophysics Data System (ADS)

Kanevski, Mikhail

2015-04-01

The research deals with an adaptation and application of Adaptive General Regression Neural Networks (GRNN) to high dimensional environmental data. GRNN [1,2,3] are efficient modelling tools both for spatial and temporal data and are based on nonparametric kernel methods closely related to classical Nadaraya-Watson estimator. Adaptive GRNN, using anisotropic kernels, can be also applied for features selection tasks when working with high dimensional data [1,3]. In the present research Adaptive GRNN are used to study geospatial data predictability and relevant feature selection using both simulated and real data case studies. The original raw data were either three dimensional monthly precipitation data or monthly wind speeds embedded into 13 dimensional space constructed by geographical coordinates and geo-features calculated from digital elevation model. GRNN were applied in two different ways: 1) adaptive GRNN with the resulting list of features ordered according to their relevancy; and 2) adaptive GRNN applied to evaluate all possible models N [in case of wind fields N=(2^13 -1)=8191] and rank them according to the cross-validation error. In both cases training were carried out applying leave-one-out procedure. An important result of the study is that the set of the most relevant features depends on the month (strong seasonal effect) and year. The predictabilities of precipitation and wind field patterns, estimated using the cross-validation and testing errors of raw and shuffled data, were studied in detail. The results of both approaches were qualitatively and quantitatively compared. In conclusion, Adaptive GRNN with their ability to select features and efficient modelling of complex high dimensional data can be widely used in automatic/on-line mapping and as an integrated part of environmental decision support systems. 1. Kanevski M., Pozdnoukhov A., Timonin V. Machine Learning for Spatial Environmental Data. Theory, applications and software. EPFL Press. With a CD: data, software, guides. (2009). 2. Kanevski M. Spatial Predictions of Soil Contamination Using General Regression Neural Networks. Systems Research and Information Systems, Volume 8, number 4, 1999. 3. Robert S., Foresti L., Kanevski M. Spatial prediction of monthly wind speeds in complex terrain with adaptive general regression neural networks. International Journal of Climatology, 33 pp. 1793-1804, 2013.

A no-reference bitstream-based perceptual model for video quality estimation of videos affected by coding artifacts and packet losses

NASA Astrophysics Data System (ADS)

Pandremmenou, K.; Shahid, M.; Kondi, L. P.; Lövström, B.

2015-03-01

In this work, we propose a No-Reference (NR) bitstream-based model for predicting the quality of H.264/AVC video sequences, affected by both compression artifacts and transmission impairments. The proposed model is based on a feature extraction procedure, where a large number of features are calculated from the packet-loss impaired bitstream. Many of the features are firstly proposed in this work, and the specific set of the features as a whole is applied for the first time for making NR video quality predictions. All feature observations are taken as input to the Least Absolute Shrinkage and Selection Operator (LASSO) regression method. LASSO indicates the most important features, and using only them, it is possible to estimate the Mean Opinion Score (MOS) with high accuracy. Indicatively, we point out that only 13 features are able to produce a Pearson Correlation Coefficient of 0.92 with the MOS. Interestingly, the performance statistics we computed in order to assess our method for predicting the Structural Similarity Index and the Video Quality Metric are equally good. Thus, the obtained experimental results verified the suitability of the features selected by LASSO as well as the ability of LASSO in making accurate predictions through sparse modeling.

Detrended fluctuation analysis for major depressive disorder.

PubMed

Mumtaz, Wajid; Malik, Aamir Saeed; Ali, Syed Saad Azhar; Yasin, Mohd Azhar Mohd; Amin, Hafeezullah

2015-01-01

Clinical utility of Electroencephalography (EEG) based diagnostic studies is less clear for major depressive disorder (MDD). In this paper, a novel machine learning (ML) scheme was presented to discriminate the MDD patients and healthy controls. The proposed method inherently involved feature extraction, selection, classification and validation. The EEG data acquisition involved eyes closed (EC) and eyes open (EO) conditions. At feature extraction stage, the de-trended fluctuation analysis (DFA) was performed, based on the EEG data, to achieve scaling exponents. The DFA was performed to analyzes the presence or absence of long-range temporal correlations (LRTC) in the recorded EEG data. The scaling exponents were used as input features to our proposed system. At feature selection stage, 3 different techniques were used for comparison purposes. Logistic regression (LR) classifier was employed. The method was validated by a 10-fold cross-validation. As results, we have observed that the effect of 3 different reference montages on the computed features. The proposed method employed 3 different types of feature selection techniques for comparison purposes as well. The results show that the DFA analysis performed better in LE data compared with the IR and AR data. In addition, during Wilcoxon ranking, the AR performed better than LE and IR. Based on the results, it was concluded that the DFA provided useful information to discriminate the MDD patients and with further validation can be employed in clinics for diagnosis of MDD.

EEG-based mild depressive detection using feature selection methods and classifiers.

PubMed

Li, Xiaowei; Hu, Bin; Sun, Shuting; Cai, Hanshu

2016-11-01

Depression has become a major health burden worldwide, and effectively detection of such disorder is a great challenge which requires latest technological tool, such as Electroencephalography (EEG). This EEG-based research seeks to find prominent frequency band and brain regions that are most related to mild depression, as well as an optimal combination of classification algorithms and feature selection methods which can be used in future mild depression detection. An experiment based on facial expression viewing task (Emo_block and Neu_block) was conducted, and EEG data of 37 university students were collected using a 128 channel HydroCel Geodesic Sensor Net (HCGSN). For discriminating mild depressive patients and normal controls, BayesNet (BN), Support Vector Machine (SVM), Logistic Regression (LR), k-nearest neighbor (KNN) and RandomForest (RF) classifiers were used. And BestFirst (BF), GreedyStepwise (GSW), GeneticSearch (GS), LinearForwordSelection (LFS) and RankSearch (RS) based on Correlation Features Selection (CFS) were applied for linear and non-linear EEG features selection. Independent Samples T-test with Bonferroni correction was used to find the significantly discriminant electrodes and features. Data mining results indicate that optimal performance is achieved using a combination of feature selection method GSW based on CFS and classifier KNN for beta frequency band. Accuracies achieved 92.00% and 98.00%, and AUC achieved 0.957 and 0.997, for Emo_block and Neu_block beta band data respectively. T-test results validate the effectiveness of selected features by search method GSW. Simplified EEG system with only FP1, FP2, F3, O2, T3 electrodes was also explored with linear features, which yielded accuracies of 91.70% and 96.00%, AUC of 0.952 and 0.972, for Emo_block and Neu_block respectively. Classification results obtained by GSW + KNN are encouraging and better than previously published results. In the spatial distribution of features, we find that left parietotemporal lobe in beta EEG frequency band has greater effect on mild depression detection. And fewer EEG channels (FP1, FP2, F3, O2 and T3) combined with linear features may be good candidates for usage in portable systems for mild depression detection. Copyright © 2016 Elsevier Ireland Ltd. All rights reserved.

Dynamic Dimensionality Selection for Bayesian Classifier Ensembles

DTIC Science & Technology

2015-03-19

learning of weights in an otherwise generatively learned naive Bayes classifier. WANBIA-C is very cometitive to Logistic Regression but much more...classifier, Generative learning, Discriminative learning, Naïve Bayes, Feature selection, Logistic regression , higher order attribute independence 16...discriminative learning of weights in an otherwise generatively learned naive Bayes classifier. WANBIA-C is very cometitive to Logistic Regression but

Comparison of naïve Bayes and logistic regression for computer-aided diagnosis of breast masses using ultrasound imaging

NASA Astrophysics Data System (ADS)

Cary, Theodore W.; Cwanger, Alyssa; Venkatesh, Santosh S.; Conant, Emily F.; Sehgal, Chandra M.

2012-03-01

This study compares the performance of two proven but very different machine learners, Naïve Bayes and logistic regression, for differentiating malignant and benign breast masses using ultrasound imaging. Ultrasound images of 266 masses were analyzed quantitatively for shape, echogenicity, margin characteristics, and texture features. These features along with patient age, race, and mammographic BI-RADS category were used to train Naïve Bayes and logistic regression classifiers to diagnose lesions as malignant or benign. ROC analysis was performed using all of the features and using only a subset that maximized information gain. Performance was determined by the area under the ROC curve, Az, obtained from leave-one-out cross validation. Naïve Bayes showed significant variation (Az 0.733 +/- 0.035 to 0.840 +/- 0.029, P < 0.002) with the choice of features, but the performance of logistic regression was relatively unchanged under feature selection (Az 0.839 +/- 0.029 to 0.859 +/- 0.028, P = 0.605). Out of 34 features, a subset of 6 gave the highest information gain: brightness difference, margin sharpness, depth-to-width, mammographic BI-RADs, age, and race. The probabilities of malignancy determined by Naïve Bayes and logistic regression after feature selection showed significant correlation (R2= 0.87, P < 0.0001). The diagnostic performance of Naïve Bayes and logistic regression can be comparable, but logistic regression is more robust. Since probability of malignancy cannot be measured directly, high correlation between the probabilities derived from two basic but dissimilar models increases confidence in the predictive power of machine learning models for characterizing solid breast masses on ultrasound.

Prediction of siRNA potency using sparse logistic regression.

PubMed

Hu, Wei; Hu, John

2014-06-01

RNA interference (RNAi) can modulate gene expression at post-transcriptional as well as transcriptional levels. Short interfering RNA (siRNA) serves as a trigger for the RNAi gene inhibition mechanism, and therefore is a crucial intermediate step in RNAi. There have been extensive studies to identify the sequence characteristics of potent siRNAs. One such study built a linear model using LASSO (Least Absolute Shrinkage and Selection Operator) to measure the contribution of each siRNA sequence feature. This model is simple and interpretable, but it requires a large number of nonzero weights. We have introduced a novel technique, sparse logistic regression, to build a linear model using single-position specific nucleotide compositions which has the same prediction accuracy of the linear model based on LASSO. The weights in our new model share the same general trend as those in the previous model, but have only 25 nonzero weights out of a total 84 weights, a 54% reduction compared to the previous model. Contrary to the linear model based on LASSO, our model suggests that only a few positions are influential on the efficacy of the siRNA, which are the 5' and 3' ends and the seed region of siRNA sequences. We also employed sparse logistic regression to build a linear model using dual-position specific nucleotide compositions, a task LASSO is not able to accomplish well due to its high dimensional nature. Our results demonstrate the superiority of sparse logistic regression as a technique for both feature selection and regression over LASSO in the context of siRNA design.

Spoken language identification based on the enhanced self-adjusting extreme learning machine approach.

PubMed

Albadr, Musatafa Abbas Abbood; Tiun, Sabrina; Al-Dhief, Fahad Taha; Sammour, Mahmoud A M

2018-01-01

Spoken Language Identification (LID) is the process of determining and classifying natural language from a given content and dataset. Typically, data must be processed to extract useful features to perform LID. The extracting features for LID, based on literature, is a mature process where the standard features for LID have already been developed using Mel-Frequency Cepstral Coefficients (MFCC), Shifted Delta Cepstral (SDC), the Gaussian Mixture Model (GMM) and ending with the i-vector based framework. However, the process of learning based on extract features remains to be improved (i.e. optimised) to capture all embedded knowledge on the extracted features. The Extreme Learning Machine (ELM) is an effective learning model used to perform classification and regression analysis and is extremely useful to train a single hidden layer neural network. Nevertheless, the learning process of this model is not entirely effective (i.e. optimised) due to the random selection of weights within the input hidden layer. In this study, the ELM is selected as a learning model for LID based on standard feature extraction. One of the optimisation approaches of ELM, the Self-Adjusting Extreme Learning Machine (SA-ELM) is selected as the benchmark and improved by altering the selection phase of the optimisation process. The selection process is performed incorporating both the Split-Ratio and K-Tournament methods, the improved SA-ELM is named Enhanced Self-Adjusting Extreme Learning Machine (ESA-ELM). The results are generated based on LID with the datasets created from eight different languages. The results of the study showed excellent superiority relating to the performance of the Enhanced Self-Adjusting Extreme Learning Machine LID (ESA-ELM LID) compared with the SA-ELM LID, with ESA-ELM LID achieving an accuracy of 96.25%, as compared to the accuracy of SA-ELM LID of only 95.00%.

Spoken language identification based on the enhanced self-adjusting extreme learning machine approach

PubMed Central

Tiun, Sabrina; AL-Dhief, Fahad Taha; Sammour, Mahmoud A. M.

2018-01-01

Spoken Language Identification (LID) is the process of determining and classifying natural language from a given content and dataset. Typically, data must be processed to extract useful features to perform LID. The extracting features for LID, based on literature, is a mature process where the standard features for LID have already been developed using Mel-Frequency Cepstral Coefficients (MFCC), Shifted Delta Cepstral (SDC), the Gaussian Mixture Model (GMM) and ending with the i-vector based framework. However, the process of learning based on extract features remains to be improved (i.e. optimised) to capture all embedded knowledge on the extracted features. The Extreme Learning Machine (ELM) is an effective learning model used to perform classification and regression analysis and is extremely useful to train a single hidden layer neural network. Nevertheless, the learning process of this model is not entirely effective (i.e. optimised) due to the random selection of weights within the input hidden layer. In this study, the ELM is selected as a learning model for LID based on standard feature extraction. One of the optimisation approaches of ELM, the Self-Adjusting Extreme Learning Machine (SA-ELM) is selected as the benchmark and improved by altering the selection phase of the optimisation process. The selection process is performed incorporating both the Split-Ratio and K-Tournament methods, the improved SA-ELM is named Enhanced Self-Adjusting Extreme Learning Machine (ESA-ELM). The results are generated based on LID with the datasets created from eight different languages. The results of the study showed excellent superiority relating to the performance of the Enhanced Self-Adjusting Extreme Learning Machine LID (ESA-ELM LID) compared with the SA-ELM LID, with ESA-ELM LID achieving an accuracy of 96.25%, as compared to the accuracy of SA-ELM LID of only 95.00%. PMID:29672546

Estimation of end point foot clearance points from inertial sensor data.

PubMed

Santhiranayagam, Braveena K; Lai, Daniel T H; Begg, Rezaul K; Palaniswami, Marimuthu

2011-01-01

Foot clearance parameters provide useful insight into tripping risks during walking. This paper proposes a technique for the estimate of key foot clearance parameters using inertial sensor (accelerometers and gyroscopes) data. Fifteen features were extracted from raw inertial sensor measurements, and a regression model was used to estimate two key foot clearance parameters: First maximum vertical clearance (m x 1) after toe-off and the Minimum Toe Clearance (MTC) of the swing foot. Comparisons are made against measurements obtained using an optoelectronic motion capture system (Optotrak), at 4 different walking speeds. General Regression Neural Networks (GRNN) were used to estimate the desired parameters from the sensor features. Eight subjects foot clearance data were examined and a Leave-one-subject-out (LOSO) method was used to select the best model. The best average Root Mean Square Errors (RMSE) across all subjects obtained using all sensor features at the maximum speed for m x 1 was 5.32 mm and for MTC was 4.04 mm. Further application of a hill-climbing feature selection technique resulted in 0.54-21.93% improvement in RMSE and required fewer input features. The results demonstrated that using raw inertial sensor data with regression models and feature selection could accurately estimate key foot clearance parameters.

Learning accurate and interpretable models based on regularized random forests regression

PubMed Central

2014-01-01

Background Many biology related research works combine data from multiple sources in an effort to understand the underlying problems. It is important to find and interpret the most important information from these sources. Thus it will be beneficial to have an effective algorithm that can simultaneously extract decision rules and select critical features for good interpretation while preserving the prediction performance. Methods In this study, we focus on regression problems for biological data where target outcomes are continuous. In general, models constructed from linear regression approaches are relatively easy to interpret. However, many practical biological applications are nonlinear in essence where we can hardly find a direct linear relationship between input and output. Nonlinear regression techniques can reveal nonlinear relationship of data, but are generally hard for human to interpret. We propose a rule based regression algorithm that uses 1-norm regularized random forests. The proposed approach simultaneously extracts a small number of rules from generated random forests and eliminates unimportant features. Results We tested the approach on some biological data sets. The proposed approach is able to construct a significantly smaller set of regression rules using a subset of attributes while achieving prediction performance comparable to that of random forests regression. Conclusion It demonstrates high potential in aiding prediction and interpretation of nonlinear relationships of the subject being studied. PMID:25350120

An Optimization-Based Method for Feature Ranking in Nonlinear Regression Problems.

PubMed

Bravi, Luca; Piccialli, Veronica; Sciandrone, Marco

2017-04-01

In this paper, we consider the feature ranking problem, where, given a set of training instances, the task is to associate a score with the features in order to assess their relevance. Feature ranking is a very important tool for decision support systems, and may be used as an auxiliary step of feature selection to reduce the high dimensionality of real-world data. We focus on regression problems by assuming that the process underlying the generated data can be approximated by a continuous function (for instance, a feedforward neural network). We formally state the notion of relevance of a feature by introducing a minimum zero-norm inversion problem of a neural network, which is a nonsmooth, constrained optimization problem. We employ a concave approximation of the zero-norm function, and we define a smooth, global optimization problem to be solved in order to assess the relevance of the features. We present the new feature ranking method based on the solution of instances of the global optimization problem depending on the available training data. Computational experiments on both artificial and real data sets are performed, and point out that the proposed feature ranking method is a valid alternative to existing methods in terms of effectiveness. The obtained results also show that the method is costly in terms of CPU time, and this may be a limitation in the solution of large-dimensional problems.

Automated retrieval of forest structure variables based on multi-scale texture analysis of VHR satellite imagery

NASA Astrophysics Data System (ADS)

Beguet, Benoit; Guyon, Dominique; Boukir, Samia; Chehata, Nesrine

2014-10-01

The main goal of this study is to design a method to describe the structure of forest stands from Very High Resolution satellite imagery, relying on some typical variables such as crown diameter, tree height, trunk diameter, tree density and tree spacing. The emphasis is placed on the automatization of the process of identification of the most relevant image features for the forest structure retrieval task, exploiting both spectral and spatial information. Our approach is based on linear regressions between the forest structure variables to be estimated and various spectral and Haralick's texture features. The main drawback of this well-known texture representation is the underlying parameters which are extremely difficult to set due to the spatial complexity of the forest structure. To tackle this major issue, an automated feature selection process is proposed which is based on statistical modeling, exploring a wide range of parameter values. It provides texture measures of diverse spatial parameters hence implicitly inducing a multi-scale texture analysis. A new feature selection technique, we called Random PRiF, is proposed. It relies on random sampling in feature space, carefully addresses the multicollinearity issue in multiple-linear regression while ensuring accurate prediction of forest variables. Our automated forest variable estimation scheme was tested on Quickbird and Pléiades panchromatic and multispectral images, acquired at different periods on the maritime pine stands of two sites in South-Western France. It outperforms two well-established variable subset selection techniques. It has been successfully applied to identify the best texture features in modeling the five considered forest structure variables. The RMSE of all predicted forest variables is improved by combining multispectral and panchromatic texture features, with various parameterizations, highlighting the potential of a multi-resolution approach for retrieving forest structure variables from VHR satellite images. Thus an average prediction error of ˜ 1.1 m is expected on crown diameter, ˜ 0.9 m on tree spacing, ˜ 3 m on height and ˜ 0.06 m on diameter at breast height.

NIMEFI: gene regulatory network inference using multiple ensemble feature importance algorithms.

PubMed

Ruyssinck, Joeri; Huynh-Thu, Vân Anh; Geurts, Pierre; Dhaene, Tom; Demeester, Piet; Saeys, Yvan

2014-01-01

One of the long-standing open challenges in computational systems biology is the topology inference of gene regulatory networks from high-throughput omics data. Recently, two community-wide efforts, DREAM4 and DREAM5, have been established to benchmark network inference techniques using gene expression measurements. In these challenges the overall top performer was the GENIE3 algorithm. This method decomposes the network inference task into separate regression problems for each gene in the network in which the expression values of a particular target gene are predicted using all other genes as possible predictors. Next, using tree-based ensemble methods, an importance measure for each predictor gene is calculated with respect to the target gene and a high feature importance is considered as putative evidence of a regulatory link existing between both genes. The contribution of this work is twofold. First, we generalize the regression decomposition strategy of GENIE3 to other feature importance methods. We compare the performance of support vector regression, the elastic net, random forest regression, symbolic regression and their ensemble variants in this setting to the original GENIE3 algorithm. To create the ensemble variants, we propose a subsampling approach which allows us to cast any feature selection algorithm that produces a feature ranking into an ensemble feature importance algorithm. We demonstrate that the ensemble setting is key to the network inference task, as only ensemble variants achieve top performance. As second contribution, we explore the effect of using rankwise averaged predictions of multiple ensemble algorithms as opposed to only one. We name this approach NIMEFI (Network Inference using Multiple Ensemble Feature Importance algorithms) and show that this approach outperforms all individual methods in general, although on a specific network a single method can perform better. An implementation of NIMEFI has been made publicly available.

Variable Selection for Regression Models of Percentile Flows

NASA Astrophysics Data System (ADS)

Fouad, G.

2017-12-01

Percentile flows describe the flow magnitude equaled or exceeded for a given percent of time, and are widely used in water resource management. However, these statistics are normally unavailable since most basins are ungauged. Percentile flows of ungauged basins are often predicted using regression models based on readily observable basin characteristics, such as mean elevation. The number of these independent variables is too large to evaluate all possible models. A subset of models is typically evaluated using automatic procedures, like stepwise regression. This ignores a large variety of methods from the field of feature (variable) selection and physical understanding of percentile flows. A study of 918 basins in the United States was conducted to compare an automatic regression procedure to the following variable selection methods: (1) principal component analysis, (2) correlation analysis, (3) random forests, (4) genetic programming, (5) Bayesian networks, and (6) physical understanding. The automatic regression procedure only performed better than principal component analysis. Poor performance of the regression procedure was due to a commonly used filter for multicollinearity, which rejected the strongest models because they had cross-correlated independent variables. Multicollinearity did not decrease model performance in validation because of a representative set of calibration basins. Variable selection methods based strictly on predictive power (numbers 2-5 from above) performed similarly, likely indicating a limit to the predictive power of the variables. Similar performance was also reached using variables selected based on physical understanding, a finding that substantiates recent calls to emphasize physical understanding in modeling for predictions in ungauged basins. The strongest variables highlighted the importance of geology and land cover, whereas widely used topographic variables were the weakest predictors. Variables suffered from a high degree of multicollinearity, possibly illustrating the co-evolution of climatic and physiographic conditions. Given the ineffectiveness of many variables used here, future work should develop new variables that target specific processes associated with percentile flows.

Alzheimer's Disease Detection by Pseudo Zernike Moment and Linear Regression Classification.

PubMed

Wang, Shui-Hua; Du, Sidan; Zhang, Yin; Phillips, Preetha; Wu, Le-Nan; Chen, Xian-Qing; Zhang, Yu-Dong

2017-01-01

This study presents an improved method based on "Gorji et al. Neuroscience. 2015" by introducing a relatively new classifier-linear regression classification. Our method selects one axial slice from 3D brain image, and employed pseudo Zernike moment with maximum order of 15 to extract 256 features from each image. Finally, linear regression classification was harnessed as the classifier. The proposed approach obtains an accuracy of 97.51%, a sensitivity of 96.71%, and a specificity of 97.73%. Our method performs better than Gorji's approach and five other state-of-the-art approaches. Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.org.

Optimizing spatial patterns with sparse filter bands for motor-imagery based brain-computer interface.

PubMed

Zhang, Yu; Zhou, Guoxu; Jin, Jing; Wang, Xingyu; Cichocki, Andrzej

2015-11-30

Common spatial pattern (CSP) has been most popularly applied to motor-imagery (MI) feature extraction for classification in brain-computer interface (BCI) application. Successful application of CSP depends on the filter band selection to a large degree. However, the most proper band is typically subject-specific and can hardly be determined manually. This study proposes a sparse filter band common spatial pattern (SFBCSP) for optimizing the spatial patterns. SFBCSP estimates CSP features on multiple signals that are filtered from raw EEG data at a set of overlapping bands. The filter bands that result in significant CSP features are then selected in a supervised way by exploiting sparse regression. A support vector machine (SVM) is implemented on the selected features for MI classification. Two public EEG datasets (BCI Competition III dataset IVa and BCI Competition IV IIb) are used to validate the proposed SFBCSP method. Experimental results demonstrate that SFBCSP help improve the classification performance of MI. The optimized spatial patterns by SFBCSP give overall better MI classification accuracy in comparison with several competing methods. The proposed SFBCSP is a potential method for improving the performance of MI-based BCI. Copyright © 2015 Elsevier B.V. All rights reserved.

Simultaneous grouping pursuit and feature selection over an undirected graph*

PubMed Central

Zhu, Yunzhang; Shen, Xiaotong; Pan, Wei

2013-01-01

Summary In high-dimensional regression, grouping pursuit and feature selection have their own merits while complementing each other in battling the curse of dimensionality. To seek a parsimonious model, we perform simultaneous grouping pursuit and feature selection over an arbitrary undirected graph with each node corresponding to one predictor. When the corresponding nodes are reachable from each other over the graph, regression coefficients can be grouped, whose absolute values are the same or close. This is motivated from gene network analysis, where genes tend to work in groups according to their biological functionalities. Through a nonconvex penalty, we develop a computational strategy and analyze the proposed method. Theoretical analysis indicates that the proposed method reconstructs the oracle estimator, that is, the unbiased least squares estimator given the true grouping, leading to consistent reconstruction of grouping structures and informative features, as well as to optimal parameter estimation. Simulation studies suggest that the method combines the benefit of grouping pursuit with that of feature selection, and compares favorably against its competitors in selection accuracy and predictive performance. An application to eQTL data is used to illustrate the methodology, where a network is incorporated into analysis through an undirected graph. PMID:24098061

Robust Joint Graph Sparse Coding for Unsupervised Spectral Feature Selection.

PubMed

Zhu, Xiaofeng; Li, Xuelong; Zhang, Shichao; Ju, Chunhua; Wu, Xindong

2017-06-01

In this paper, we propose a new unsupervised spectral feature selection model by embedding a graph regularizer into the framework of joint sparse regression for preserving the local structures of data. To do this, we first extract the bases of training data by previous dictionary learning methods and, then, map original data into the basis space to generate their new representations, by proposing a novel joint graph sparse coding (JGSC) model. In JGSC, we first formulate its objective function by simultaneously taking subspace learning and joint sparse regression into account, then, design a new optimization solution to solve the resulting objective function, and further prove the convergence of the proposed solution. Furthermore, we extend JGSC to a robust JGSC (RJGSC) via replacing the least square loss function with a robust loss function, for achieving the same goals and also avoiding the impact of outliers. Finally, experimental results on real data sets showed that both JGSC and RJGSC outperformed the state-of-the-art algorithms in terms of k -nearest neighbor classification performance.

Comparing Mammography Abnormality Features and Genetic Variants in the Prediction of Breast Cancer in Women Recommended for Breast Biopsy

PubMed Central

Burnside, Elizabeth S.; Liu, Jie; Wu, Yirong; Onitilo, Adedayo A.; McCarty, Catherine; Page, C. David; Peissig, Peggy; Trentham-Dietz, Amy; Kitchner, Terrie; Fan, Jun; Yuan, Ming

2015-01-01

Rationale and Objectives The discovery of germline genetic variants associated with breast cancer has engendered interest in risk stratification for improved, targeted detection and diagnosis. However, there has yet to be a comparison of the predictive ability of these genetic variants with mammography abnormality descriptors. Materials and Methods Our IRB-approved, HIPAA-compliant study utilized a personalized medicine registry in which participants consented to provide a DNA sample and participate in longitudinal follow-up. In our retrospective, age-matched, case-controlled study of 373 cases and 395 controls who underwent breast biopsy, we collected risk factors selected a priori based on the literature including: demographic variables based on the Gail model, common germline genetic variants, and diagnostic mammography findings according to BI-RADS. We developed predictive models using logistic regression to determine the predictive ability of: 1) demographic variables, 2) 10 selected genetic variants, or 3) mammography BI-RADS features. We evaluated each model in turn by calculating a risk score for each patient using 10-fold cross validation; used this risk estimate to construct ROC curves; and compared the AUC of each using the DeLong method. Results The performance of the regression model using demographic risk factors was not statistically different from the model using genetic variants (p=0.9). The model using mammography features (AUC = 0.689) was superior to both the demographic model (AUC = .598; p<0.001) and the genetic model (AUC = .601; p<0.001). Conclusion BI-RADS features exceeded the ability of demographic and 10 selected germline genetic variants to predict breast cancer in women recommended for biopsy. PMID:26514439

OSA severity assessment based on sleep breathing analysis using ambient microphone.

PubMed

Dafna, E; Tarasiuk, A; Zigel, Y

2013-01-01

In this paper, an audio-based system for severity estimation of obstructive sleep apnea (OSA) is proposed. The system estimates the apnea-hypopnea index (AHI), which is the average number of apneic events per hour of sleep. This system is based on a Gaussian mixture regression algorithm that was trained and validated on full-night audio recordings. Feature selection process using a genetic algorithm was applied to select the best features extracted from time and spectra domains. A total of 155 subjects, referred to in-laboratory polysomnography (PSG) study, were recruited. Using the PSG's AHI score as a gold-standard, the performances of the proposed system were evaluated using a Pearson correlation, AHI error, and diagnostic agreement methods. Correlation of R=0.89, AHI error of 7.35 events/hr, and diagnostic agreement of 77.3% were achieved, showing encouraging performances and a reliable non-contact alternative method for OSA severity estimation.

SU-F-R-20: Image Texture Features Correlate with Time to Local Failure in Lung SBRT Patients

DOE Office of Scientific and Technical Information (OSTI.GOV)

Andrews, M; Abazeed, M; Woody, N

Purpose: To explore possible correlation between CT image-based texture and histogram features and time-to-local-failure in early stage non-small cell lung cancer (NSCLC) patients treated with stereotactic body radiotherapy (SBRT).Methods and Materials: From an IRB-approved lung SBRT registry for patients treated between 2009–2013 we selected 48 (20 male, 28 female) patients with local failure. Median patient age was 72.3±10.3 years. Mean time to local failure was 15 ± 7.1 months. Physician-contoured gross tumor volumes (GTV) on the planning CT images were processed and 3D gray-level co-occurrence matrix (GLCM) based texture and histogram features were calculated in Matlab. Data were exported tomore » R and a multiple linear regression model was used to examine the relationship between texture features and time-to-local-failure. Results: Multiple linear regression revealed that entropy (p=0.0233, multiple R2=0.60) from GLCM-based texture analysis and the standard deviation (p=0.0194, multiple R2=0.60) from the histogram-based features were statistically significantly correlated with the time-to-local-failure. Conclusion: Image-based texture analysis can be used to predict certain aspects of treatment outcomes of NSCLC patients treated with SBRT. We found entropy and standard deviation calculated for the GTV on the CT images displayed a statistically significant correlation with and time-to-local-failure in lung SBRT patients.« less

Determining Cutoff Point of Ensemble Trees Based on Sample Size in Predicting Clinical Dose with DNA Microarray Data.

PubMed

Yılmaz Isıkhan, Selen; Karabulut, Erdem; Alpar, Celal Reha

2016-01-01

Background/Aim . Evaluating the success of dose prediction based on genetic or clinical data has substantially advanced recently. The aim of this study is to predict various clinical dose values from DNA gene expression datasets using data mining techniques. Materials and Methods . Eleven real gene expression datasets containing dose values were included. First, important genes for dose prediction were selected using iterative sure independence screening. Then, the performances of regression trees (RTs), support vector regression (SVR), RT bagging, SVR bagging, and RT boosting were examined. Results . The results demonstrated that a regression-based feature selection method substantially reduced the number of irrelevant genes from raw datasets. Overall, the best prediction performance in nine of 11 datasets was achieved using SVR; the second most accurate performance was provided using a gradient-boosting machine (GBM). Conclusion . Analysis of various dose values based on microarray gene expression data identified common genes found in our study and the referenced studies. According to our findings, SVR and GBM can be good predictors of dose-gene datasets. Another result of the study was to identify the sample size of n = 25 as a cutoff point for RT bagging to outperform a single RT.

Cyst-based measurements for assessing lymphangioleiomyomatosis in computed tomography

DOE Office of Scientific and Technical Information (OSTI.GOV)

Lo, P., E-mail: pechinlo@mednet.edu.ucla; Brown, M. S.; Kim, H.

Purpose: To investigate the efficacy of a new family of measurements made on individual pulmonary cysts extracted from computed tomography (CT) for assessing the severity of lymphangioleiomyomatosis (LAM). Methods: CT images were analyzed using thresholding to identify a cystic region of interest from chest CT of LAM patients. Individual cysts were then extracted from the cystic region by the watershed algorithm, which separates individual cysts based on subtle edges within the cystic regions. A family of measurements were then computed, which quantify the amount, distribution, and boundary appearance of the cysts. Sequential floating feature selection was used to select amore » small subset of features for quantification of the severity of LAM. Adjusted R{sup 2} from multiple linear regression and R{sup 2} from linear regression against measurements from spirometry were used to compare the performance of our proposed measurements with currently used density based CT measurements in the literature, namely, the relative area measure and the D measure. Results: Volumetric CT data, performed at total lung capacity and residual volume, from a total of 49 subjects enrolled in the MILES trial were used in our study. Our proposed measures had adjusted R{sup 2} ranging from 0.42 to 0.59 when regressing against the spirometry measures, with p < 0.05. For previously used density based CT measurements in the literature, the best R{sup 2} was 0.46 (for only one instance), with the majority being lower than 0.3 or p > 0.05. Conclusions: The proposed family of CT-based cyst measurements have better correlation with spirometric measures than previously used density based CT measurements. They show potential as a sensitive tool for quantitatively assessing the severity of LAM.« less

Strategies for minimizing sample size for use in airborne LiDAR-based forest inventory

USGS Publications Warehouse

Junttila, Virpi; Finley, Andrew O.; Bradford, John B.; Kauranne, Tuomo

2013-01-01

Recently airborne Light Detection And Ranging (LiDAR) has emerged as a highly accurate remote sensing modality to be used in operational scale forest inventories. Inventories conducted with the help of LiDAR are most often model-based, i.e. they use variables derived from LiDAR point clouds as the predictive variables that are to be calibrated using field plots. The measurement of the necessary field plots is a time-consuming and statistically sensitive process. Because of this, current practice often presumes hundreds of plots to be collected. But since these plots are only used to calibrate regression models, it should be possible to minimize the number of plots needed by carefully selecting the plots to be measured. In the current study, we compare several systematic and random methods for calibration plot selection, with the specific aim that they be used in LiDAR based regression models for forest parameters, especially above-ground biomass. The primary criteria compared are based on both spatial representativity as well as on their coverage of the variability of the forest features measured. In the former case, it is important also to take into account spatial auto-correlation between the plots. The results indicate that choosing the plots in a way that ensures ample coverage of both spatial and feature space variability improves the performance of the corresponding models, and that adequate coverage of the variability in the feature space is the most important condition that should be met by the set of plots collected.

Morphologic Features of Magnetic Resonance Imaging as a Surrogate of Capsular Contracture in Breast Cancer Patients With Implant-based Reconstructions.

PubMed

Tyagi, Neelam; Sutton, Elizabeth; Hunt, Margie; Zhang, Jing; Oh, Jung Hun; Apte, Aditya; Mechalakos, James; Wilgucki, Molly; Gelb, Emily; Mehrara, Babak; Matros, Evan; Ho, Alice

2017-02-01

Capsular contracture (CC) is a serious complication in patients receiving implant-based reconstruction for breast cancer. Currently, no objective methods are available for assessing CC. The goal of the present study was to identify image-based surrogates of CC using magnetic resonance imaging (MRI). We analyzed a retrospective data set of 50 patients who had undergone both a diagnostic MRI scan and a plastic surgeon's evaluation of the CC score (Baker's score) within a 6-month period after mastectomy and reconstructive surgery. The MRI scans were assessed for morphologic shape features of the implant and histogram features of the pectoralis muscle. The shape features, such as roundness, eccentricity, solidity, extent, and ratio length for the implant, were compared with the Baker score. For the pectoralis muscle, the muscle width and median, skewness, and kurtosis of the intensity were compared with the Baker score. Univariate analysis (UVA) using a Wilcoxon rank-sum test and multivariate analysis with the least absolute shrinkage and selection operator logistic regression was performed to determine significant differences in these features between the patient groups categorized according to their Baker's scores. UVA showed statistically significant differences between grade 1 and grade ≥2 for morphologic shape features and histogram features, except for volume and skewness. Only eccentricity, ratio length, and volume were borderline significant in differentiating grade ≤2 and grade ≥3. Features with P<.1 on UVA were used in the multivariate least absolute shrinkage and selection operator logistic regression analysis. Multivariate analysis showed a good level of predictive power for grade 1 versus grade ≥2 CC (area under the receiver operating characteristic curve 0.78, sensitivity 0.78, and specificity 0.82) and for grade ≤2 versus grade ≥3 CC (area under the receiver operating characteristic curve 0.75, sensitivity 0.75, and specificity 0.79). The morphologic shape features described on MR images were associated with the severity of CC. MRI has the potential to further improve the diagnostic ability of the Baker score in breast cancer patients who undergo implant reconstruction. Copyright © 2016 Elsevier Inc. All rights reserved.

Improving Lidar-based Aboveground Biomass Estimation with Site Productivity for Central Hardwood Forests, USA

NASA Astrophysics Data System (ADS)

Shao, G.; Gallion, J.; Fei, S.

2016-12-01

Sound forest aboveground biomass estimation is required to monitor diverse forest ecosystems and their impacts on the changing climate. Lidar-based regression models provided promised biomass estimations in most forest ecosystems. However, considerable uncertainties of biomass estimations have been reported in the temperate hardwood and hardwood-dominated mixed forests. Varied site productivities in temperate hardwood forests largely diversified height and diameter growth rates, which significantly reduced the correlation between tree height and diameter at breast height (DBH) in mature and complex forests. It is, therefore, difficult to utilize height-based lidar metrics to predict DBH-based field-measured biomass through a simple regression model regardless the variation of site productivity. In this study, we established a multi-dimension nonlinear regression model incorporating lidar metrics and site productivity classes derived from soil features. In the regression model, lidar metrics provided horizontal and vertical structural information and productivity classes differentiated good and poor forest sites. The selection and combination of lidar metrics were discussed. Multiple regression models were employed and compared. Uncertainty analysis was applied to the best fit model. The effects of site productivity on the lidar-based biomass model were addressed.

TH-E-BRF-05: Comparison of Survival-Time Prediction Models After Radiotherapy for High-Grade Glioma Patients Based On Clinical and DVH Features

DOE Office of Scientific and Technical Information (OSTI.GOV)

Magome, T; Haga, A; Igaki, H

Purpose: Although many outcome prediction models based on dose-volume information have been proposed, it is well known that the prognosis may be affected also by multiple clinical factors. The purpose of this study is to predict the survival time after radiotherapy for high-grade glioma patients based on features including clinical and dose-volume histogram (DVH) information. Methods: A total of 35 patients with high-grade glioma (oligodendroglioma: 2, anaplastic astrocytoma: 3, glioblastoma: 30) were selected in this study. All patients were treated with prescribed dose of 30–80 Gy after surgical resection or biopsy from 2006 to 2013 at The University of Tokyomore » Hospital. All cases were randomly separated into training dataset (30 cases) and test dataset (5 cases). The survival time after radiotherapy was predicted based on a multiple linear regression analysis and artificial neural network (ANN) by using 204 candidate features. The candidate features included the 12 clinical features (tumor location, extent of surgical resection, treatment duration of radiotherapy, etc.), and the 192 DVH features (maximum dose, minimum dose, D95, V60, etc.). The effective features for the prediction were selected according to a step-wise method by using 30 training cases. The prediction accuracy was evaluated by a coefficient of determination (R{sup 2}) between the predicted and actual survival time for the training and test dataset. Results: In the multiple regression analysis, the value of R{sup 2} between the predicted and actual survival time was 0.460 for the training dataset and 0.375 for the test dataset. On the other hand, in the ANN analysis, the value of R{sup 2} was 0.806 for the training dataset and 0.811 for the test dataset. Conclusion: Although a large number of patients would be needed for more accurate and robust prediction, our preliminary Result showed the potential to predict the outcome in the patients with high-grade glioma. This work was partly supported by the JSPS Core-to-Core Program(No. 23003) and Grant-in-aid from the JSPS Fellows.« less

Comparison of l₁-Norm SVR and Sparse Coding Algorithms for Linear Regression.

PubMed

Zhang, Qingtian; Hu, Xiaolin; Zhang, Bo

2015-08-01

Support vector regression (SVR) is a popular function estimation technique based on Vapnik's concept of support vector machine. Among many variants, the l1-norm SVR is known to be good at selecting useful features when the features are redundant. Sparse coding (SC) is a technique widely used in many areas and a number of efficient algorithms are available. Both l1-norm SVR and SC can be used for linear regression. In this brief, the close connection between the l1-norm SVR and SC is revealed and some typical algorithms are compared for linear regression. The results show that the SC algorithms outperform the Newton linear programming algorithm, an efficient l1-norm SVR algorithm, in efficiency. The algorithms are then used to design the radial basis function (RBF) neural networks. Experiments on some benchmark data sets demonstrate the high efficiency of the SC algorithms. In particular, one of the SC algorithms, the orthogonal matching pursuit is two orders of magnitude faster than a well-known RBF network designing algorithm, the orthogonal least squares algorithm.

NIMEFI: Gene Regulatory Network Inference using Multiple Ensemble Feature Importance Algorithms

PubMed Central

Ruyssinck, Joeri; Huynh-Thu, Vân Anh; Geurts, Pierre; Dhaene, Tom; Demeester, Piet; Saeys, Yvan

2014-01-01

One of the long-standing open challenges in computational systems biology is the topology inference of gene regulatory networks from high-throughput omics data. Recently, two community-wide efforts, DREAM4 and DREAM5, have been established to benchmark network inference techniques using gene expression measurements. In these challenges the overall top performer was the GENIE3 algorithm. This method decomposes the network inference task into separate regression problems for each gene in the network in which the expression values of a particular target gene are predicted using all other genes as possible predictors. Next, using tree-based ensemble methods, an importance measure for each predictor gene is calculated with respect to the target gene and a high feature importance is considered as putative evidence of a regulatory link existing between both genes. The contribution of this work is twofold. First, we generalize the regression decomposition strategy of GENIE3 to other feature importance methods. We compare the performance of support vector regression, the elastic net, random forest regression, symbolic regression and their ensemble variants in this setting to the original GENIE3 algorithm. To create the ensemble variants, we propose a subsampling approach which allows us to cast any feature selection algorithm that produces a feature ranking into an ensemble feature importance algorithm. We demonstrate that the ensemble setting is key to the network inference task, as only ensemble variants achieve top performance. As second contribution, we explore the effect of using rankwise averaged predictions of multiple ensemble algorithms as opposed to only one. We name this approach NIMEFI (Network Inference using Multiple Ensemble Feature Importance algorithms) and show that this approach outperforms all individual methods in general, although on a specific network a single method can perform better. An implementation of NIMEFI has been made publicly available. PMID:24667482

Ultrasound based computer-aided-diagnosis of kidneys for pediatric hydronephrosis

NASA Astrophysics Data System (ADS)

Cerrolaza, Juan J.; Peters, Craig A.; Martin, Aaron D.; Myers, Emmarie; Safdar, Nabile; Linguraru, Marius G.

2014-03-01

Ultrasound is the mainstay of imaging for pediatric hydronephrosis, though its potential as diagnostic tool is limited by its subjective assessment, and lack of correlation with renal function. Therefore, all cases showing signs of hydronephrosis undergo further invasive studies, like diuretic renogram, in order to assess the actual renal function. Under the hypothesis that renal morphology is correlated with renal function, a new ultrasound based computer-aided diagnosis (CAD) tool for pediatric hydronephrosis is presented. From 2D ultrasound, a novel set of morphological features of the renal collecting systems and the parenchyma, is automatically extracted using image analysis techniques. From the original set of features, including size, geometric and curvature descriptors, a subset of ten features are selected as predictive variables, combining a feature selection technique and area under the curve filtering. Using the washout half time (T1/2) as indicative of renal obstruction, two groups are defined. Those cases whose T1/2 is above 30 minutes are considered to be severe, while the rest would be in the safety zone, where diuretic renography could be avoided. Two different classification techniques are evaluated (logistic regression, and support vector machines). Adjusting the probability decision thresholds to operate at the point of maximum sensitivity, i.e., preventing any severe case be misclassified, specificities of 53%, and 75% are achieved, for the logistic regression and the support vector machine classifier, respectively. The proposed CAD system allows to establish a link between non-invasive non-ionizing imaging techniques and renal function, limiting the need for invasive and ionizing diuretic renography.

Quality optimization of H.264/AVC video transmission over noisy environments using a sparse regression framework

NASA Astrophysics Data System (ADS)

Pandremmenou, K.; Tziortziotis, N.; Paluri, S.; Zhang, W.; Blekas, K.; Kondi, L. P.; Kumar, S.

2015-03-01

We propose the use of the Least Absolute Shrinkage and Selection Operator (LASSO) regression method in order to predict the Cumulative Mean Squared Error (CMSE), incurred by the loss of individual slices in video transmission. We extract a number of quality-relevant features from the H.264/AVC video sequences, which are given as input to the LASSO. This method has the benefit of not only keeping a subset of the features that have the strongest effects towards video quality, but also produces accurate CMSE predictions. Particularly, we study the LASSO regression through two different architectures; the Global LASSO (G.LASSO) and Local LASSO (L.LASSO). In G.LASSO, a single regression model is trained for all slice types together, while in L.LASSO, motivated by the fact that the values for some features are closely dependent on the considered slice type, each slice type has its own regression model, in an e ort to improve LASSO's prediction capability. Based on the predicted CMSE values, we group the video slices into four priority classes. Additionally, we consider a video transmission scenario over a noisy channel, where Unequal Error Protection (UEP) is applied to all prioritized slices. The provided results demonstrate the efficiency of LASSO in estimating CMSE with high accuracy, using only a few features. les that typically contain high-entropy data, producing a footprint that is far less conspicuous than existing methods. The system uses a local web server to provide a le system, user interface and applications through an web architecture.

A wavelet-based technique to predict treatment outcome for Major Depressive Disorder.

PubMed

Mumtaz, Wajid; Xia, Likun; Mohd Yasin, Mohd Azhar; Azhar Ali, Syed Saad; Malik, Aamir Saeed

2017-01-01

Treatment management for Major Depressive Disorder (MDD) has been challenging. However, electroencephalogram (EEG)-based predictions of antidepressant's treatment outcome may help during antidepressant's selection and ultimately improve the quality of life for MDD patients. In this study, a machine learning (ML) method involving pretreatment EEG data was proposed to perform such predictions for Selective Serotonin Reuptake Inhibitor (SSRIs). For this purpose, the acquisition of experimental data involved 34 MDD patients and 30 healthy controls. Consequently, a feature matrix was constructed involving time-frequency decomposition of EEG data based on wavelet transform (WT) analysis, termed as EEG data matrix. However, the resultant EEG data matrix had high dimensionality. Therefore, dimension reduction was performed based on a rank-based feature selection method according to a criterion, i.e., receiver operating characteristic (ROC). As a result, the most significant features were identified and further be utilized during the training and testing of a classification model, i.e., the logistic regression (LR) classifier. Finally, the LR model was validated with 100 iterations of 10-fold cross-validation (10-CV). The classification results were compared with short-time Fourier transform (STFT) analysis, and empirical mode decompositions (EMD). The wavelet features extracted from frontal and temporal EEG data were found statistically significant. In comparison with other time-frequency approaches such as the STFT and EMD, the WT analysis has shown highest classification accuracy, i.e., accuracy = 87.5%, sensitivity = 95%, and specificity = 80%. In conclusion, significant wavelet coefficients extracted from frontal and temporal pre-treatment EEG data involving delta and theta frequency bands may predict antidepressant's treatment outcome for the MDD patients.

A Cross-Domain Collaborative Filtering Algorithm Based on Feature Construction and Locally Weighted Linear Regression

PubMed Central

Jiang, Feng; Han, Ji-zhong

2018-01-01

Cross-domain collaborative filtering (CDCF) solves the sparsity problem by transferring rating knowledge from auxiliary domains. Obviously, different auxiliary domains have different importance to the target domain. However, previous works cannot evaluate effectively the significance of different auxiliary domains. To overcome this drawback, we propose a cross-domain collaborative filtering algorithm based on Feature Construction and Locally Weighted Linear Regression (FCLWLR). We first construct features in different domains and use these features to represent different auxiliary domains. Thus the weight computation across different domains can be converted as the weight computation across different features. Then we combine the features in the target domain and in the auxiliary domains together and convert the cross-domain recommendation problem into a regression problem. Finally, we employ a Locally Weighted Linear Regression (LWLR) model to solve the regression problem. As LWLR is a nonparametric regression method, it can effectively avoid underfitting or overfitting problem occurring in parametric regression methods. We conduct extensive experiments to show that the proposed FCLWLR algorithm is effective in addressing the data sparsity problem by transferring the useful knowledge from the auxiliary domains, as compared to many state-of-the-art single-domain or cross-domain CF methods. PMID:29623088

A Cross-Domain Collaborative Filtering Algorithm Based on Feature Construction and Locally Weighted Linear Regression.

PubMed

Yu, Xu; Lin, Jun-Yu; Jiang, Feng; Du, Jun-Wei; Han, Ji-Zhong

2018-01-01

Cross-domain collaborative filtering (CDCF) solves the sparsity problem by transferring rating knowledge from auxiliary domains. Obviously, different auxiliary domains have different importance to the target domain. However, previous works cannot evaluate effectively the significance of different auxiliary domains. To overcome this drawback, we propose a cross-domain collaborative filtering algorithm based on Feature Construction and Locally Weighted Linear Regression (FCLWLR). We first construct features in different domains and use these features to represent different auxiliary domains. Thus the weight computation across different domains can be converted as the weight computation across different features. Then we combine the features in the target domain and in the auxiliary domains together and convert the cross-domain recommendation problem into a regression problem. Finally, we employ a Locally Weighted Linear Regression (LWLR) model to solve the regression problem. As LWLR is a nonparametric regression method, it can effectively avoid underfitting or overfitting problem occurring in parametric regression methods. We conduct extensive experiments to show that the proposed FCLWLR algorithm is effective in addressing the data sparsity problem by transferring the useful knowledge from the auxiliary domains, as compared to many state-of-the-art single-domain or cross-domain CF methods.

Connectome-based predictive modeling of attention: Comparing different functional connectivity features and prediction methods across datasets.

PubMed

Yoo, Kwangsun; Rosenberg, Monica D; Hsu, Wei-Ting; Zhang, Sheng; Li, Chiang-Shan R; Scheinost, Dustin; Constable, R Todd; Chun, Marvin M

2018-02-15

Connectome-based predictive modeling (CPM; Finn et al., 2015; Shen et al., 2017) was recently developed to predict individual differences in traits and behaviors, including fluid intelligence (Finn et al., 2015) and sustained attention (Rosenberg et al., 2016a), from functional brain connectivity (FC) measured with fMRI. Here, using the CPM framework, we compared the predictive power of three different measures of FC (Pearson's correlation, accordance, and discordance) and two different prediction algorithms (linear and partial least square [PLS] regression) for attention function. Accordance and discordance are recently proposed FC measures that respectively track in-phase synchronization and out-of-phase anti-correlation (Meskaldji et al., 2015). We defined connectome-based models using task-based or resting-state FC data, and tested the effects of (1) functional connectivity measure and (2) feature-selection/prediction algorithm on individualized attention predictions. Models were internally validated in a training dataset using leave-one-subject-out cross-validation, and externally validated with three independent datasets. The training dataset included fMRI data collected while participants performed a sustained attention task and rested (N = 25; Rosenberg et al., 2016a). The validation datasets included: 1) data collected during performance of a stop-signal task and at rest (N = 83, including 19 participants who were administered methylphenidate prior to scanning; Farr et al., 2014a; Rosenberg et al., 2016b), 2) data collected during Attention Network Task performance and rest (N = 41, Rosenberg et al., in press), and 3) resting-state data and ADHD symptom severity from the ADHD-200 Consortium (N = 113; Rosenberg et al., 2016a). Models defined using all combinations of functional connectivity measure (Pearson's correlation, accordance, and discordance) and prediction algorithm (linear and PLS regression) predicted attentional abilities, with correlations between predicted and observed measures of attention as high as 0.9 for internal validation, and 0.6 for external validation (all p's < 0.05). Models trained on task data outperformed models trained on rest data. Pearson's correlation and accordance features generally showed a small numerical advantage over discordance features, while PLS regression models were usually better than linear regression models. Overall, in addition to correlation features combined with linear models (Rosenberg et al., 2016a), it is useful to consider accordance features and PLS regression for CPM. Copyright © 2017 Elsevier Inc. All rights reserved.

Minimalist ensemble algorithms for genome-wide protein localization prediction.

PubMed

Lin, Jhih-Rong; Mondal, Ananda Mohan; Liu, Rong; Hu, Jianjun

2012-07-03

Computational prediction of protein subcellular localization can greatly help to elucidate its functions. Despite the existence of dozens of protein localization prediction algorithms, the prediction accuracy and coverage are still low. Several ensemble algorithms have been proposed to improve the prediction performance, which usually include as many as 10 or more individual localization algorithms. However, their performance is still limited by the running complexity and redundancy among individual prediction algorithms. This paper proposed a novel method for rational design of minimalist ensemble algorithms for practical genome-wide protein subcellular localization prediction. The algorithm is based on combining a feature selection based filter and a logistic regression classifier. Using a novel concept of contribution scores, we analyzed issues of algorithm redundancy, consensus mistakes, and algorithm complementarity in designing ensemble algorithms. We applied the proposed minimalist logistic regression (LR) ensemble algorithm to two genome-wide datasets of Yeast and Human and compared its performance with current ensemble algorithms. Experimental results showed that the minimalist ensemble algorithm can achieve high prediction accuracy with only 1/3 to 1/2 of individual predictors of current ensemble algorithms, which greatly reduces computational complexity and running time. It was found that the high performance ensemble algorithms are usually composed of the predictors that together cover most of available features. Compared to the best individual predictor, our ensemble algorithm improved the prediction accuracy from AUC score of 0.558 to 0.707 for the Yeast dataset and from 0.628 to 0.646 for the Human dataset. Compared with popular weighted voting based ensemble algorithms, our classifier-based ensemble algorithms achieved much better performance without suffering from inclusion of too many individual predictors. We proposed a method for rational design of minimalist ensemble algorithms using feature selection and classifiers. The proposed minimalist ensemble algorithm based on logistic regression can achieve equal or better prediction performance while using only half or one-third of individual predictors compared to other ensemble algorithms. The results also suggested that meta-predictors that take advantage of a variety of features by combining individual predictors tend to achieve the best performance. The LR ensemble server and related benchmark datasets are available at http://mleg.cse.sc.edu/LRensemble/cgi-bin/predict.cgi.

Minimalist ensemble algorithms for genome-wide protein localization prediction

PubMed Central

2012-01-01

Background Computational prediction of protein subcellular localization can greatly help to elucidate its functions. Despite the existence of dozens of protein localization prediction algorithms, the prediction accuracy and coverage are still low. Several ensemble algorithms have been proposed to improve the prediction performance, which usually include as many as 10 or more individual localization algorithms. However, their performance is still limited by the running complexity and redundancy among individual prediction algorithms. Results This paper proposed a novel method for rational design of minimalist ensemble algorithms for practical genome-wide protein subcellular localization prediction. The algorithm is based on combining a feature selection based filter and a logistic regression classifier. Using a novel concept of contribution scores, we analyzed issues of algorithm redundancy, consensus mistakes, and algorithm complementarity in designing ensemble algorithms. We applied the proposed minimalist logistic regression (LR) ensemble algorithm to two genome-wide datasets of Yeast and Human and compared its performance with current ensemble algorithms. Experimental results showed that the minimalist ensemble algorithm can achieve high prediction accuracy with only 1/3 to 1/2 of individual predictors of current ensemble algorithms, which greatly reduces computational complexity and running time. It was found that the high performance ensemble algorithms are usually composed of the predictors that together cover most of available features. Compared to the best individual predictor, our ensemble algorithm improved the prediction accuracy from AUC score of 0.558 to 0.707 for the Yeast dataset and from 0.628 to 0.646 for the Human dataset. Compared with popular weighted voting based ensemble algorithms, our classifier-based ensemble algorithms achieved much better performance without suffering from inclusion of too many individual predictors. Conclusions We proposed a method for rational design of minimalist ensemble algorithms using feature selection and classifiers. The proposed minimalist ensemble algorithm based on logistic regression can achieve equal or better prediction performance while using only half or one-third of individual predictors compared to other ensemble algorithms. The results also suggested that meta-predictors that take advantage of a variety of features by combining individual predictors tend to achieve the best performance. The LR ensemble server and related benchmark datasets are available at http://mleg.cse.sc.edu/LRensemble/cgi-bin/predict.cgi. PMID:22759391

Characterization of electroencephalography signals for estimating saliency features in videos.

PubMed

Liang, Zhen; Hamada, Yasuyuki; Oba, Shigeyuki; Ishii, Shin

2018-05-12

Understanding the functions of the visual system has been one of the major targets in neuroscience formany years. However, the relation between spontaneous brain activities and visual saliency in natural stimuli has yet to be elucidated. In this study, we developed an optimized machine learning-based decoding model to explore the possible relationships between the electroencephalography (EEG) characteristics and visual saliency. The optimal features were extracted from the EEG signals and saliency map which was computed according to an unsupervised saliency model ( Tavakoli and Laaksonen, 2017). Subsequently, various unsupervised feature selection/extraction techniques were examined using different supervised regression models. The robustness of the presented model was fully verified by means of ten-fold or nested cross validation procedure, and promising results were achieved in the reconstruction of saliency features based on the selected EEG characteristics. Through the successful demonstration of using EEG characteristics to predict the real-time saliency distribution in natural videos, we suggest the feasibility of quantifying visual content through measuring brain activities (EEG signals) in real environments, which would facilitate the understanding of cortical involvement in the processing of natural visual stimuli and application developments motivated by human visual processing. Copyright © 2018 Elsevier Ltd. All rights reserved.

Heterogeneity wavelet kinetics from DCE-MRI for classifying gene expression based breast cancer recurrence risk.

PubMed

Mahrooghy, Majid; Ashraf, Ahmed B; Daye, Dania; Mies, Carolyn; Feldman, Michael; Rosen, Mark; Kontos, Despina

2013-01-01

Breast tumors are heterogeneous lesions. Intra-tumor heterogeneity presents a major challenge for cancer diagnosis and treatment. Few studies have worked on capturing tumor heterogeneity from imaging. Most studies to date consider aggregate measures for tumor characterization. In this work we capture tumor heterogeneity by partitioning tumor pixels into subregions and extracting heterogeneity wavelet kinetic (HetWave) features from breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to obtain the spatiotemporal patterns of the wavelet coefficients and contrast agent uptake from each partition. Using a genetic algorithm for feature selection, and a logistic regression classifier with leave one-out cross validation, we tested our proposed HetWave features for the task of classifying breast cancer recurrence risk. The classifier based on our features gave an ROC AUC of 0.78, outperforming previously proposed kinetic, texture, and spatial enhancement variance features which give AUCs of 0.69, 0.64, and 0.65, respectively.

New strategy for determination of anthocyanins, polyphenols and antioxidant capacity of Brassica oleracea liquid extract using infrared spectroscopies and multivariate regression

NASA Astrophysics Data System (ADS)

de Oliveira, Isadora R. N.; Roque, Jussara V.; Maia, Mariza P.; Stringheta, Paulo C.; Teófilo, Reinaldo F.

2018-04-01

A new method was developed to determine the antioxidant properties of red cabbage extract (Brassica oleracea) by mid (MID) and near (NIR) infrared spectroscopies and partial least squares (PLS) regression. A 70% (v/v) ethanolic extract of red cabbage was concentrated to 9° Brix and further diluted (12 to 100%) in water. The dilutions were used as external standards for the building of PLS models. For the first time, this strategy was applied for building multivariate regression models. Reference analyses and spectral data were obtained from diluted extracts. The determinate properties were total and monomeric anthocyanins, total polyphenols and antioxidant capacity by ABTS (2,2-azino-bis(3-ethyl-benzothiazoline-6-sulfonate)) and DPPH (2,2-diphenyl-1-picrylhydrazyl) methods. Ordered predictors selection (OPS) and genetic algorithm (GA) were used for feature selection before PLS regression (PLS-1). In addition, a PLS-2 regression was applied to all properties simultaneously. PLS-1 models provided more predictive models than did PLS-2 regression. PLS-OPS and PLS-GA models presented excellent prediction results with a correlation coefficient higher than 0.98. However, the best models were obtained using PLS and variable selection with the OPS algorithm and the models based on NIR spectra were considered more predictive for all properties. Then, these models provided a simple, rapid and accurate method for determination of red cabbage extract antioxidant properties and its suitability for use in the food industry.

Computer-aided Classification of Mammographic Masses Using Visually Sensitive Image Features

PubMed Central

Wang, Yunzhi; Aghaei, Faranak; Zarafshani, Ali; Qiu, Yuchen; Qian, Wei; Zheng, Bin

2017-01-01

Purpose To develop a new computer-aided diagnosis (CAD) scheme that computes visually sensitive image features routinely used by radiologists to develop a machine learning classifier and distinguish between the malignant and benign breast masses detected from digital mammograms. Methods An image dataset including 301 breast masses was retrospectively selected. From each segmented mass region, we computed image features that mimic five categories of visually sensitive features routinely used by radiologists in reading mammograms. We then selected five optimal features in the five feature categories and applied logistic regression models for classification. A new CAD interface was also designed to show lesion segmentation, computed feature values and classification score. Results Areas under ROC curves (AUC) were 0.786±0.026 and 0.758±0.027 when to classify mass regions depicting on two view images, respectively. By fusing classification scores computed from two regions, AUC increased to 0.806±0.025. Conclusion This study demonstrated a new approach to develop CAD scheme based on 5 visually sensitive image features. Combining with a “visual aid” interface, CAD results may be much more easily explainable to the observers and increase their confidence to consider CAD generated classification results than using other conventional CAD approaches, which involve many complicated and visually insensitive texture features. PMID:27911353

Analysing Twitter and web queries for flu trend prediction.

PubMed

Santos, José Carlos; Matos, Sérgio

2014-05-07

Social media platforms encourage people to share diverse aspects of their daily life. Among these, shared health related information might be used to infer health status and incidence rates for specific conditions or symptoms. In this work, we present an infodemiology study that evaluates the use of Twitter messages and search engine query logs to estimate and predict the incidence rate of influenza like illness in Portugal. Based on a manually classified dataset of 2704 tweets from Portugal, we selected a set of 650 textual features to train a Naïve Bayes classifier to identify tweets mentioning flu or flu-like illness or symptoms. We obtained a precision of 0.78 and an F-measure of 0.83, based on cross validation over the complete annotated set. Furthermore, we trained a multiple linear regression model to estimate the health-monitoring data from the Influenzanet project, using as predictors the relative frequencies obtained from the tweet classification results and from query logs, and achieved a correlation ratio of 0.89 (p<0.001). These classification and regression models were also applied to estimate the flu incidence in the following flu season, achieving a correlation of 0.72. Previous studies addressing the estimation of disease incidence based on user-generated content have mostly focused on the english language. Our results further validate those studies and show that by changing the initial steps of data preprocessing and feature extraction and selection, the proposed approaches can be adapted to other languages. Additionally, we investigated whether the predictive model created can be applied to data from the subsequent flu season. In this case, although the prediction result was good, an initial phase to adapt the regression model could be necessary to achieve more robust results.

WE-E-17A-02: Predictive Modeling of Outcome Following SABR for NSCLC Based On Radiomics of FDG-PET Images

DOE Office of Scientific and Technical Information (OSTI.GOV)

Li, R; Aguilera, T; Shultz, D

2014-06-15

Purpose: This study aims to develop predictive models of patient outcome by extracting advanced imaging features (i.e., Radiomics) from FDG-PET images. Methods: We acquired pre-treatment PET scans for 51 stage I NSCLC patients treated with SABR. We calculated 139 quantitative features from each patient PET image, including 5 morphological features, 8 statistical features, 27 texture features, and 100 features from the intensity-volume histogram. Based on the imaging features, we aim to distinguish between 2 risk groups of patients: those with regional failure or distant metastasis versus those without. We investigated 3 pattern classification algorithms: linear discriminant analysis (LDA), naive Bayesmore » (NB), and logistic regression (LR). To avoid the curse of dimensionality, we performed feature selection by first removing redundant features and then applying sequential forward selection using the wrapper approach. To evaluate the predictive performance, we performed 10-fold cross validation with 1000 random splits of the data and calculated the area under the ROC curve (AUC). Results: Feature selection identified 2 texture features (homogeneity and/or wavelet decompositions) for NB and LR, while for LDA SUVmax and one texture feature (correlation) were identified. All 3 classifiers achieved statistically significant improvements over conventional PET imaging metrics such as tumor volume (AUC = 0.668) and SUVmax (AUC = 0.737). Overall, NB achieved the best predictive performance (AUC = 0.806). This also compares favorably with MTV using the best threshold at an SUV of 11.6 (AUC = 0.746). At a sensitivity of 80%, NB achieved 69% specificity, while SUVmax and tumor volume only had 36% and 47% specificity. Conclusion: Through a systematic analysis of advanced PET imaging features, we are able to build models with improved predictive value over conventional imaging metrics. If validated in a large independent cohort, the proposed techniques could potentially aid in identifying patients who might benefit from adjuvant therapy.« less

Computing group cardinality constraint solutions for logistic regression problems.

PubMed

Zhang, Yong; Kwon, Dongjin; Pohl, Kilian M

2017-01-01

We derive an algorithm to directly solve logistic regression based on cardinality constraint, group sparsity and use it to classify intra-subject MRI sequences (e.g. cine MRIs) of healthy from diseased subjects. Group cardinality constraint models are often applied to medical images in order to avoid overfitting of the classifier to the training data. Solutions within these models are generally determined by relaxing the cardinality constraint to a weighted feature selection scheme. However, these solutions relate to the original sparse problem only under specific assumptions, which generally do not hold for medical image applications. In addition, inferring clinical meaning from features weighted by a classifier is an ongoing topic of discussion. Avoiding weighing features, we propose to directly solve the group cardinality constraint logistic regression problem by generalizing the Penalty Decomposition method. To do so, we assume that an intra-subject series of images represents repeated samples of the same disease patterns. We model this assumption by combining series of measurements created by a feature across time into a single group. Our algorithm then derives a solution within that model by decoupling the minimization of the logistic regression function from enforcing the group sparsity constraint. The minimum to the smooth and convex logistic regression problem is determined via gradient descent while we derive a closed form solution for finding a sparse approximation of that minimum. We apply our method to cine MRI of 38 healthy controls and 44 adult patients that received reconstructive surgery of Tetralogy of Fallot (TOF) during infancy. Our method correctly identifies regions impacted by TOF and generally obtains statistically significant higher classification accuracy than alternative solutions to this model, i.e., ones relaxing group cardinality constraints. Copyright © 2016 Elsevier B.V. All rights reserved.

Harmony Search as a Powerful Tool for Feature Selection in QSPR Study of the Drugs Lipophilicity.

PubMed

Bahadori, Behnoosh; Atabati, Morteza

2017-01-01

Aims & Scope: Lipophilicity represents one of the most studied and most frequently used fundamental physicochemical properties. In the present work, harmony search (HS) algorithm is suggested to feature selection in quantitative structure-property relationship (QSPR) modeling to predict lipophilicity of neutral, acidic, basic and amphotheric drugs that were determined by UHPLC. Harmony search is a music-based metaheuristic optimization algorithm. It was affected by the observation that the aim of music is to search for a perfect state of harmony. Semi-empirical quantum-chemical calculations at AM1 level were used to find the optimum 3D geometry of the studied molecules and variant descriptors (1497 descriptors) were calculated by the Dragon software. The selected descriptors by harmony search algorithm (9 descriptors) were applied for model development using multiple linear regression (MLR). In comparison with other feature selection methods such as genetic algorithm and simulated annealing, harmony search algorithm has better results. The root mean square error (RMSE) with and without leave-one out cross validation (LOOCV) were obtained 0.417 and 0.302, respectively. The results were compared with those obtained from the genetic algorithm and simulated annealing methods and it showed that the HS is a helpful tool for feature selection with fine performance. Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.org.

Interictal epileptiform discharge characteristics underlying expert interrater agreement.

PubMed

Bagheri, Elham; Dauwels, Justin; Dean, Brian C; Waters, Chad G; Westover, M Brandon; Halford, Jonathan J

2017-10-01

The presence of interictal epileptiform discharges (IED) in the electroencephalogram (EEG) is a key finding in the medical workup of a patient with suspected epilepsy. However, inter-rater agreement (IRA) regarding the presence of IED is imperfect, leading to incorrect and delayed diagnoses. An improved understanding of which IED attributes mediate expert IRA might help in developing automatic methods for IED detection able to emulate the abilities of experts. Therefore, using a set of IED scored by a large number of experts, we set out to determine which attributes of IED predict expert agreement regarding the presence of IED. IED were annotated on a 5-point scale by 18 clinical neurophysiologists within 200 30-s EEG segments from recordings of 200 patients. 5538 signal analysis features were extracted from the waveforms, including wavelet coefficients, morphological features, signal energy, nonlinear energy operator response, electrode location, and spectrogram features. Feature selection was performed by applying elastic net regression and support vector regression (SVR) was applied to predict expert opinion, with and without the feature selection procedure and with and without several types of signal normalization. Multiple types of features were useful for predicting expert annotations, but particular types of wavelet features performed best. Local EEG normalization also enhanced best model performance. As the size of the group of EEGers used to train the models was increased, the performance of the models leveled off at a group size of around 11. The features that best predict inter-rater agreement among experts regarding the presence of IED are wavelet features, using locally standardized EEG. Our models for predicting expert opinion based on EEGer's scores perform best with a large group of EEGers (more than 10). By examining a large group of EEG signal analysis features we found that wavelet features with certain wavelet basis functions performed best to identify IEDs. Local normalization also improves predictability, suggesting the importance of IED morphology over amplitude-based features. Although most IED detection studies in the past have used opinion from three or fewer experts, our study suggests a "wisdom of the crowd" effect, such that pooling over a larger number of expert opinions produces a better correlation between expert opinion and objectively quantifiable features of the EEG. Copyright © 2017 International Federation of Clinical Neurophysiology. Published by Elsevier B.V. All rights reserved.

Computational Intelligence Modeling of the Macromolecules Release from PLGA Microspheres-Focus on Feature Selection.

PubMed

Zawbaa, Hossam M; Szlȩk, Jakub; Grosan, Crina; Jachowicz, Renata; Mendyk, Aleksander

2016-01-01

Poly-lactide-co-glycolide (PLGA) is a copolymer of lactic and glycolic acid. Drug release from PLGA microspheres depends not only on polymer properties but also on drug type, particle size, morphology of microspheres, release conditions, etc. Selecting a subset of relevant properties for PLGA is a challenging machine learning task as there are over three hundred features to consider. In this work, we formulate the selection of critical attributes for PLGA as a multiobjective optimization problem with the aim of minimizing the error of predicting the dissolution profile while reducing the number of attributes selected. Four bio-inspired optimization algorithms: antlion optimization, binary version of antlion optimization, grey wolf optimization, and social spider optimization are used to select the optimal feature set for predicting the dissolution profile of PLGA. Besides these, LASSO algorithm is also used for comparisons. Selection of crucial variables is performed under the assumption that both predictability and model simplicity are of equal importance to the final result. During the feature selection process, a set of input variables is employed to find minimum generalization error across different predictive models and their settings/architectures. The methodology is evaluated using predictive modeling for which various tools are chosen, such as Cubist, random forests, artificial neural networks (monotonic MLP, deep learning MLP), multivariate adaptive regression splines, classification and regression tree, and hybrid systems of fuzzy logic and evolutionary computations (fugeR). The experimental results are compared with the results reported by Szlȩk. We obtain a normalized root mean square error (NRMSE) of 15.97% versus 15.4%, and the number of selected input features is smaller, nine versus eleven.

Computational Intelligence Modeling of the Macromolecules Release from PLGA Microspheres—Focus on Feature Selection

PubMed Central

Zawbaa, Hossam M.; Szlȩk, Jakub; Grosan, Crina; Jachowicz, Renata; Mendyk, Aleksander

2016-01-01

Poly-lactide-co-glycolide (PLGA) is a copolymer of lactic and glycolic acid. Drug release from PLGA microspheres depends not only on polymer properties but also on drug type, particle size, morphology of microspheres, release conditions, etc. Selecting a subset of relevant properties for PLGA is a challenging machine learning task as there are over three hundred features to consider. In this work, we formulate the selection of critical attributes for PLGA as a multiobjective optimization problem with the aim of minimizing the error of predicting the dissolution profile while reducing the number of attributes selected. Four bio-inspired optimization algorithms: antlion optimization, binary version of antlion optimization, grey wolf optimization, and social spider optimization are used to select the optimal feature set for predicting the dissolution profile of PLGA. Besides these, LASSO algorithm is also used for comparisons. Selection of crucial variables is performed under the assumption that both predictability and model simplicity are of equal importance to the final result. During the feature selection process, a set of input variables is employed to find minimum generalization error across different predictive models and their settings/architectures. The methodology is evaluated using predictive modeling for which various tools are chosen, such as Cubist, random forests, artificial neural networks (monotonic MLP, deep learning MLP), multivariate adaptive regression splines, classification and regression tree, and hybrid systems of fuzzy logic and evolutionary computations (fugeR). The experimental results are compared with the results reported by Szlȩk. We obtain a normalized root mean square error (NRMSE) of 15.97% versus 15.4%, and the number of selected input features is smaller, nine versus eleven. PMID:27315205

A wavelet-based technique to predict treatment outcome for Major Depressive Disorder

PubMed Central

Xia, Likun; Mohd Yasin, Mohd Azhar; Azhar Ali, Syed Saad

2017-01-01

Treatment management for Major Depressive Disorder (MDD) has been challenging. However, electroencephalogram (EEG)-based predictions of antidepressant’s treatment outcome may help during antidepressant’s selection and ultimately improve the quality of life for MDD patients. In this study, a machine learning (ML) method involving pretreatment EEG data was proposed to perform such predictions for Selective Serotonin Reuptake Inhibitor (SSRIs). For this purpose, the acquisition of experimental data involved 34 MDD patients and 30 healthy controls. Consequently, a feature matrix was constructed involving time-frequency decomposition of EEG data based on wavelet transform (WT) analysis, termed as EEG data matrix. However, the resultant EEG data matrix had high dimensionality. Therefore, dimension reduction was performed based on a rank-based feature selection method according to a criterion, i.e., receiver operating characteristic (ROC). As a result, the most significant features were identified and further be utilized during the training and testing of a classification model, i.e., the logistic regression (LR) classifier. Finally, the LR model was validated with 100 iterations of 10-fold cross-validation (10-CV). The classification results were compared with short-time Fourier transform (STFT) analysis, and empirical mode decompositions (EMD). The wavelet features extracted from frontal and temporal EEG data were found statistically significant. In comparison with other time-frequency approaches such as the STFT and EMD, the WT analysis has shown highest classification accuracy, i.e., accuracy = 87.5%, sensitivity = 95%, and specificity = 80%. In conclusion, significant wavelet coefficients extracted from frontal and temporal pre-treatment EEG data involving delta and theta frequency bands may predict antidepressant’s treatment outcome for the MDD patients. PMID:28152063

Real estate value prediction using multivariate regression models

NASA Astrophysics Data System (ADS)

Manjula, R.; Jain, Shubham; Srivastava, Sharad; Rajiv Kher, Pranav

2017-11-01

The real estate market is one of the most competitive in terms of pricing and the same tends to vary significantly based on a lot of factors, hence it becomes one of the prime fields to apply the concepts of machine learning to optimize and predict the prices with high accuracy. Therefore in this paper, we present various important features to use while predicting housing prices with good accuracy. We have described regression models, using various features to have lower Residual Sum of Squares error. While using features in a regression model some feature engineering is required for better prediction. Often a set of features (multiple regressions) or polynomial regression (applying a various set of powers in the features) is used for making better model fit. For these models are expected to be susceptible towards over fitting ridge regression is used to reduce it. This paper thus directs to the best application of regression models in addition to other techniques to optimize the result.

Algorithm For Solution Of Subset-Regression Problems

NASA Technical Reports Server (NTRS)

Verhaegen, Michel

1991-01-01

Reliable and flexible algorithm for solution of subset-regression problem performs QR decomposition with new column-pivoting strategy, enables selection of subset directly from originally defined regression parameters. This feature, in combination with number of extensions, makes algorithm very flexible for use in analysis of subset-regression problems in which parameters have physical meanings. Also extended to enable joint processing of columns contaminated by noise with those free of noise, without using scaling techniques.

EEG feature selection method based on decision tree.

PubMed

Duan, Lijuan; Ge, Hui; Ma, Wei; Miao, Jun

2015-01-01

This paper aims to solve automated feature selection problem in brain computer interface (BCI). In order to automate feature selection process, we proposed a novel EEG feature selection method based on decision tree (DT). During the electroencephalogram (EEG) signal processing, a feature extraction method based on principle component analysis (PCA) was used, and the selection process based on decision tree was performed by searching the feature space and automatically selecting optimal features. Considering that EEG signals are a series of non-linear signals, a generalized linear classifier named support vector machine (SVM) was chosen. In order to test the validity of the proposed method, we applied the EEG feature selection method based on decision tree to BCI Competition II datasets Ia, and the experiment showed encouraging results.

Cascaded face alignment via intimacy definition feature

NASA Astrophysics Data System (ADS)

Li, Hailiang; Lam, Kin-Man; Chiu, Man-Yau; Wu, Kangheng; Lei, Zhibin

2017-09-01

Recent years have witnessed the emerging popularity of regression-based face aligners, which directly learn mappings between facial appearance and shape-increment manifolds. We propose a random-forest based, cascaded regression model for face alignment by using a locally lightweight feature, namely intimacy definition feature. This feature is more discriminative than the pose-indexed feature, more efficient than the histogram of oriented gradients feature and the scale-invariant feature transform feature, and more compact than the local binary feature (LBF). Experimental validation of our algorithm shows that our approach achieves state-of-the-art performance when testing on some challenging datasets. Compared with the LBF-based algorithm, our method achieves about twice the speed, 20% improvement in terms of alignment accuracy and saves an order of magnitude on memory requirement.

A Robust Shape Reconstruction Method for Facial Feature Point Detection.

PubMed

Tan, Shuqiu; Chen, Dongyi; Guo, Chenggang; Huang, Zhiqi

2017-01-01

Facial feature point detection has been receiving great research advances in recent years. Numerous methods have been developed and applied in practical face analysis systems. However, it is still a quite challenging task because of the large variability in expression and gestures and the existence of occlusions in real-world photo shoot. In this paper, we present a robust sparse reconstruction method for the face alignment problems. Instead of a direct regression between the feature space and the shape space, the concept of shape increment reconstruction is introduced. Moreover, a set of coupled overcomplete dictionaries termed the shape increment dictionary and the local appearance dictionary are learned in a regressive manner to select robust features and fit shape increments. Additionally, to make the learned model more generalized, we select the best matched parameter set through extensive validation tests. Experimental results on three public datasets demonstrate that the proposed method achieves a better robustness over the state-of-the-art methods.

Habitat features and predictive habitat modeling for the Colorado chipmunk in southern New Mexico

USGS Publications Warehouse

Rivieccio, M.; Thompson, B.C.; Gould, W.R.; Boykin, K.G.

2003-01-01

Two subspecies of Colorado chipmunk (state threatened and federal species of concern) occur in southern New Mexico: Tamias quadrivittatus australis in the Organ Mountains and T. q. oscuraensis in the Oscura Mountains. We developed a GIS model of potentially suitable habitat based on vegetation and elevation features, evaluated site classifications of the GIS model, and determined vegetation and terrain features associated with chipmunk occurrence. We compared GIS model classifications with actual vegetation and elevation features measured at 37 sites. At 60 sites we measured 18 habitat variables regarding slope, aspect, tree species, shrub species, and ground cover. We used logistic regression to analyze habitat variables associated with chipmunk presence/absence. All (100%) 37 sample sites (28 predicted suitable, 9 predicted unsuitable) were classified correctly by the GIS model regarding elevation and vegetation. For 28 sites predicted suitable by the GIS model, 18 sites (64%) appeared visually suitable based on habitat variables selected from logistic regression analyses, of which 10 sites (36%) were specifically predicted as suitable habitat via logistic regression. We detected chipmunks at 70% of sites deemed suitable via the logistic regression models. Shrub cover, tree density, plant proximity, presence of logs, and presence of rock outcrop were retained in the logistic model for the Oscura Mountains; litter, shrub cover, and grass cover were retained in the logistic model for the Organ Mountains. Evaluation of predictive models illustrates the need for multi-stage analyses to best judge performance. Microhabitat analyses indicate prospective needs for different management strategies between the subspecies. Sensitivities of each population of the Colorado chipmunk to natural and prescribed fire suggest that partial burnings of areas inhabited by Colorado chipmunks in southern New Mexico may be beneficial. These partial burnings may later help avoid a fire that could substantially reduce habitat of chipmunks over a mountain range.

Feature Selection for Motor Imagery EEG Classification Based on Firefly Algorithm and Learning Automata

PubMed Central

Liu, Aiming; Liu, Quan; Ai, Qingsong; Xie, Yi; Chen, Anqi

2017-01-01

Motor Imagery (MI) electroencephalography (EEG) is widely studied for its non-invasiveness, easy availability, portability, and high temporal resolution. As for MI EEG signal processing, the high dimensions of features represent a research challenge. It is necessary to eliminate redundant features, which not only create an additional overhead of managing the space complexity, but also might include outliers, thereby reducing classification accuracy. The firefly algorithm (FA) can adaptively select the best subset of features, and improve classification accuracy. However, the FA is easily entrapped in a local optimum. To solve this problem, this paper proposes a method of combining the firefly algorithm and learning automata (LA) to optimize feature selection for motor imagery EEG. We employed a method of combining common spatial pattern (CSP) and local characteristic-scale decomposition (LCD) algorithms to obtain a high dimensional feature set, and classified it by using the spectral regression discriminant analysis (SRDA) classifier. Both the fourth brain–computer interface competition data and real-time data acquired in our designed experiments were used to verify the validation of the proposed method. Compared with genetic and adaptive weight particle swarm optimization algorithms, the experimental results show that our proposed method effectively eliminates redundant features, and improves the classification accuracy of MI EEG signals. In addition, a real-time brain–computer interface system was implemented to verify the feasibility of our proposed methods being applied in practical brain–computer interface systems. PMID:29117100

Feature Selection for Motor Imagery EEG Classification Based on Firefly Algorithm and Learning Automata.

PubMed

Liu, Aiming; Chen, Kun; Liu, Quan; Ai, Qingsong; Xie, Yi; Chen, Anqi

2017-11-08

Motor Imagery (MI) electroencephalography (EEG) is widely studied for its non-invasiveness, easy availability, portability, and high temporal resolution. As for MI EEG signal processing, the high dimensions of features represent a research challenge. It is necessary to eliminate redundant features, which not only create an additional overhead of managing the space complexity, but also might include outliers, thereby reducing classification accuracy. The firefly algorithm (FA) can adaptively select the best subset of features, and improve classification accuracy. However, the FA is easily entrapped in a local optimum. To solve this problem, this paper proposes a method of combining the firefly algorithm and learning automata (LA) to optimize feature selection for motor imagery EEG. We employed a method of combining common spatial pattern (CSP) and local characteristic-scale decomposition (LCD) algorithms to obtain a high dimensional feature set, and classified it by using the spectral regression discriminant analysis (SRDA) classifier. Both the fourth brain-computer interface competition data and real-time data acquired in our designed experiments were used to verify the validation of the proposed method. Compared with genetic and adaptive weight particle swarm optimization algorithms, the experimental results show that our proposed method effectively eliminates redundant features, and improves the classification accuracy of MI EEG signals. In addition, a real-time brain-computer interface system was implemented to verify the feasibility of our proposed methods being applied in practical brain-computer interface systems.

Identification of eggs from different production systems based on hyperspectra and CS-SVM.

PubMed

Sun, J; Cong, S L; Mao, H P; Zhou, X; Wu, X H; Zhang, X D

2017-06-01

1. To identify the origin of table eggs more accurately, a method based on hyperspectral imaging technology was studied. 2. The hyperspectral data of 200 samples of intensive and extensive eggs were collected. Standard normalised variables combined with a Savitzky-Golay were used to eliminate noise, then stepwise regression (SWR) was used for feature selection. Grid search algorithm (GS), genetic search algorithm (GA), particle swarm optimisation algorithm (PSO) and cuckoo search algorithm (CS) were applied by support vector machine (SVM) methods to establish an SVM identification model with the optimal parameters. The full spectrum data and the data after feature selection were the input of the model, while egg category was the output. 3. The SWR-CS-SVM model performed better than the other models, including SWR-GS-SVM, SWR-GA-SVM, SWR-PSO-SVM and others based on full spectral data. The training and test classification accuracy of the SWR-CS-SVM model were respectively 99.3% and 96%. 4. SWR-CS-SVM proved effective for identifying egg varieties and could also be useful for the non-destructive identification of other types of egg.

Toward better public health reporting using existing off the shelf approaches: A comparison of alternative cancer detection approaches using plaintext medical data and non-dictionary based feature selection.

PubMed

Kasthurirathne, Suranga N; Dixon, Brian E; Gichoya, Judy; Xu, Huiping; Xia, Yuni; Mamlin, Burke; Grannis, Shaun J

2016-04-01

Increased adoption of electronic health records has resulted in increased availability of free text clinical data for secondary use. A variety of approaches to obtain actionable information from unstructured free text data exist. These approaches are resource intensive, inherently complex and rely on structured clinical data and dictionary-based approaches. We sought to evaluate the potential to obtain actionable information from free text pathology reports using routinely available tools and approaches that do not depend on dictionary-based approaches. We obtained pathology reports from a large health information exchange and evaluated the capacity to detect cancer cases from these reports using 3 non-dictionary feature selection approaches, 4 feature subset sizes, and 5 clinical decision models: simple logistic regression, naïve bayes, k-nearest neighbor, random forest, and J48 decision tree. The performance of each decision model was evaluated using sensitivity, specificity, accuracy, positive predictive value, and area under the receiver operating characteristics (ROC) curve. Decision models parameterized using automated, informed, and manual feature selection approaches yielded similar results. Furthermore, non-dictionary classification approaches identified cancer cases present in free text reports with evaluation measures approaching and exceeding 80-90% for most metrics. Our methods are feasible and practical approaches for extracting substantial information value from free text medical data, and the results suggest that these methods can perform on par, if not better, than existing dictionary-based approaches. Given that public health agencies are often under-resourced and lack the technical capacity for more complex methodologies, these results represent potentially significant value to the public health field. Copyright © 2016 Elsevier Inc. All rights reserved.

Variable Selection for Road Segmentation in Aerial Images

NASA Astrophysics Data System (ADS)

Warnke, S.; Bulatov, D.

2017-05-01

For extraction of road pixels from combined image and elevation data, Wegner et al. (2015) proposed classification of superpixels into road and non-road, after which a refinement of the classification results using minimum cost paths and non-local optimization methods took place. We believed that the variable set used for classification was to a certain extent suboptimal, because many variables were redundant while several features known as useful in Photogrammetry and Remote Sensing are missed. This motivated us to implement a variable selection approach which builds a model for classification using portions of training data and subsets of features, evaluates this model, updates the feature set, and terminates when a stopping criterion is satisfied. The choice of classifier is flexible; however, we tested the approach with Logistic Regression and Random Forests, and taylored the evaluation module to the chosen classifier. To guarantee a fair comparison, we kept the segment-based approach and most of the variables from the related work, but we extended them by additional, mostly higher-level features. Applying these superior features, removing the redundant ones, as well as using more accurately acquired 3D data allowed to keep stable or even to reduce the misclassification error in a challenging dataset.

Near-infrared spectral image analysis of pork marbling based on Gabor filter and wide line detector techniques.

PubMed

Huang, Hui; Liu, Li; Ngadi, Michael O; Gariépy, Claude; Prasher, Shiv O

2014-01-01

Marbling is an important quality attribute of pork. Detection of pork marbling usually involves subjective scoring, which raises the efficiency costs to the processor. In this study, the ability to predict pork marbling using near-infrared (NIR) hyperspectral imaging (900-1700 nm) and the proper image processing techniques were studied. Near-infrared images were collected from pork after marbling evaluation according to current standard chart from the National Pork Producers Council. Image analysis techniques-Gabor filter, wide line detector, and spectral averaging-were applied to extract texture, line, and spectral features, respectively, from NIR images of pork. Samples were grouped into calibration and validation sets. Wavelength selection was performed on calibration set by stepwise regression procedure. Prediction models of pork marbling scores were built using multiple linear regressions based on derivatives of mean spectra and line features at key wavelengths. The results showed that the derivatives of both texture and spectral features produced good results, with correlation coefficients of validation of 0.90 and 0.86, respectively, using wavelengths of 961, 1186, and 1220 nm. The results revealed the great potential of the Gabor filter for analyzing NIR images of pork for the effective and efficient objective evaluation of pork marbling.

A Novel Continuous Blood Pressure Estimation Approach Based on Data Mining Techniques.

PubMed

Miao, Fen; Fu, Nan; Zhang, Yuan-Ting; Ding, Xiao-Rong; Hong, Xi; He, Qingyun; Li, Ye

2017-11-01

Continuous blood pressure (BP) estimation using pulse transit time (PTT) is a promising method for unobtrusive BP measurement. However, the accuracy of this approach must be improved for it to be viable for a wide range of applications. This study proposes a novel continuous BP estimation approach that combines data mining techniques with a traditional mechanism-driven model. First, 14 features derived from simultaneous electrocardiogram and photoplethysmogram signals were extracted for beat-to-beat BP estimation. A genetic algorithm-based feature selection method was then used to select BP indicators for each subject. Multivariate linear regression and support vector regression were employed to develop the BP model. The accuracy and robustness of the proposed approach were validated for static, dynamic, and follow-up performance. Experimental results based on 73 subjects showed that the proposed approach exhibited excellent accuracy in static BP estimation, with a correlation coefficient and mean error of 0.852 and -0.001 ± 3.102 mmHg for systolic BP, and 0.790 and -0.004 ± 2.199 mmHg for diastolic BP. Similar performance was observed for dynamic BP estimation. The robustness results indicated that the estimation accuracy was lower by a certain degree one day after model construction but was relatively stable from one day to six months after construction. The proposed approach is superior to the state-of-the-art PTT-based model for an approximately 2-mmHg reduction in the standard derivation at different time intervals, thus providing potentially novel insights for cuffless BP estimation.

LINKING LUNG AIRWAY STRUCTURE TO PULMONARY FUNCTION VIA COMPOSITE BRIDGE REGRESSION

PubMed Central

Chen, Kun; Hoffman, Eric A.; Seetharaman, Indu; Jiao, Feiran; Lin, Ching-Long; Chan, Kung-Sik

2017-01-01

The human lung airway is a complex inverted tree-like structure. Detailed airway measurements can be extracted from MDCT-scanned lung images, such as segmental wall thickness, airway diameter, parent-child branch angles, etc. The wealth of lung airway data provides a unique opportunity for advancing our understanding of the fundamental structure-function relationships within the lung. An important problem is to construct and identify important lung airway features in normal subjects and connect these to standardized pulmonary function test results such as FEV1%. Among other things, the problem is complicated by the fact that a particular airway feature may be an important (relevant) predictor only when it pertains to segments of certain generations. Thus, the key is an efficient, consistent method for simultaneously conducting group selection (lung airway feature types) and within-group variable selection (airway generations), i.e., bi-level selection. Here we streamline a comprehensive procedure to process the lung airway data via imputation, normalization, transformation and groupwise principal component analysis, and then adopt a new composite penalized regression approach for conducting bi-level feature selection. As a prototype of composite penalization, the proposed composite bridge regression method is shown to admit an efficient algorithm, enjoy bi-level oracle properties, and outperform several existing methods. We analyze the MDCT lung image data from a cohort of 132 subjects with normal lung function. Our results show that, lung function in terms of FEV1% is promoted by having a less dense and more homogeneous lung comprising an airway whose segments enjoy more heterogeneity in wall thicknesses, larger mean diameters, lumen areas and branch angles. These data hold the potential of defining more accurately the “normal” subject population with borderline atypical lung functions that are clearly influenced by many genetic and environmental factors. PMID:28280520

Integrated feature extraction and selection for neuroimage classification

NASA Astrophysics Data System (ADS)

Fan, Yong; Shen, Dinggang

2009-02-01

Feature extraction and selection are of great importance in neuroimage classification for identifying informative features and reducing feature dimensionality, which are generally implemented as two separate steps. This paper presents an integrated feature extraction and selection algorithm with two iterative steps: constrained subspace learning based feature extraction and support vector machine (SVM) based feature selection. The subspace learning based feature extraction focuses on the brain regions with higher possibility of being affected by the disease under study, while the possibility of brain regions being affected by disease is estimated by the SVM based feature selection, in conjunction with SVM classification. This algorithm can not only take into account the inter-correlation among different brain regions, but also overcome the limitation of traditional subspace learning based feature extraction methods. To achieve robust performance and optimal selection of parameters involved in feature extraction, selection, and classification, a bootstrapping strategy is used to generate multiple versions of training and testing sets for parameter optimization, according to the classification performance measured by the area under the ROC (receiver operating characteristic) curve. The integrated feature extraction and selection method is applied to a structural MR image based Alzheimer's disease (AD) study with 98 non-demented and 100 demented subjects. Cross-validation results indicate that the proposed algorithm can improve performance of the traditional subspace learning based classification.

Association between mammogram density and background parenchymal enhancement of breast MRI

NASA Astrophysics Data System (ADS)

Aghaei, Faranak; Danala, Gopichandh; Wang, Yunzhi; Zarafshani, Ali; Qian, Wei; Liu, Hong; Zheng, Bin

2018-02-01

Breast density has been widely considered as an important risk factor for breast cancer. The purpose of this study is to examine the association between mammogram density results and background parenchymal enhancement (BPE) of breast MRI. A dataset involving breast MR images was acquired from 65 high-risk women. Based on mammography density (BIRADS) results, the dataset was divided into two groups of low and high breast density cases. The Low-Density group has 15 cases with mammographic density (BIRADS 1 and 2), while the High-density group includes 50 cases, which were rated by radiologists as mammographic density BIRADS 3 and 4. A computer-aided detection (CAD) scheme was applied to segment and register breast regions depicted on sequential images of breast MRI scans. CAD scheme computed 20 global BPE features from the entire two breast regions, separately from the left and right breast region, as well as from the bilateral difference between left and right breast regions. An image feature selection method namely, CFS method, was applied to remove the most redundant features and select optimal features from the initial feature pool. Then, a logistic regression classifier was built using the optimal features to predict the mammogram density from the BPE features. Using a leave-one-case-out validation method, the classifier yields the accuracy of 82% and area under ROC curve, AUC=0.81+/-0.09. Also, the box-plot based analysis shows a negative association between mammogram density results and BPE features in the MRI images. This study demonstrated a negative association between mammogram density and BPE of breast MRI images.

Economic indicators selection for crime rates forecasting using cooperative feature selection

NASA Astrophysics Data System (ADS)

Alwee, Razana; Shamsuddin, Siti Mariyam Hj; Salleh Sallehuddin, Roselina

2013-04-01

Features selection in multivariate forecasting model is very important to ensure that the model is accurate. The purpose of this study is to apply the Cooperative Feature Selection method for features selection. The features are economic indicators that will be used in crime rate forecasting model. The Cooperative Feature Selection combines grey relational analysis and artificial neural network to establish a cooperative model that can rank and select the significant economic indicators. Grey relational analysis is used to select the best data series to represent each economic indicator and is also used to rank the economic indicators according to its importance to the crime rate. After that, the artificial neural network is used to select the significant economic indicators for forecasting the crime rates. In this study, we used economic indicators of unemployment rate, consumer price index, gross domestic product and consumer sentiment index, as well as data rates of property crime and violent crime for the United States. Levenberg-Marquardt neural network is used in this study. From our experiments, we found that consumer price index is an important economic indicator that has a significant influence on the violent crime rate. While for property crime rate, the gross domestic product, unemployment rate and consumer price index are the influential economic indicators. The Cooperative Feature Selection is also found to produce smaller errors as compared to Multiple Linear Regression in forecasting property and violent crime rates.

Spectral Regression Based Fault Feature Extraction for Bearing Accelerometer Sensor Signals

PubMed Central

Xia, Zhanguo; Xia, Shixiong; Wan, Ling; Cai, Shiyu

2012-01-01

Bearings are not only the most important element but also a common source of failures in rotary machinery. Bearing fault prognosis technology has been receiving more and more attention recently, in particular because it plays an increasingly important role in avoiding the occurrence of accidents. Therein, fault feature extraction (FFE) of bearing accelerometer sensor signals is essential to highlight representative features of bearing conditions for machinery fault diagnosis and prognosis. This paper proposes a spectral regression (SR)-based approach for fault feature extraction from original features including time, frequency and time-frequency domain features of bearing accelerometer sensor signals. SR is a novel regression framework for efficient regularized subspace learning and feature extraction technology, and it uses the least squares method to obtain the best projection direction, rather than computing the density matrix of features, so it also has the advantage in dimensionality reduction. The effectiveness of the SR-based method is validated experimentally by applying the acquired vibration signals data to bearings. The experimental results indicate that SR can reduce the computation cost and preserve more structure information about different bearing faults and severities, and it is demonstrated that the proposed feature extraction scheme has an advantage over other similar approaches. PMID:23202017

An EEG-based functional connectivity measure for automatic detection of alcohol use disorder.

PubMed

Mumtaz, Wajid; Saad, Mohamad Naufal B Mohamad; Kamel, Nidal; Ali, Syed Saad Azhar; Malik, Aamir Saeed

2018-01-01

The abnormal alcohol consumption could cause toxicity and could alter the human brain's structure and function, termed as alcohol used disorder (AUD). Unfortunately, the conventional screening methods for AUD patients are subjective and manual. Hence, to perform automatic screening of AUD patients, objective methods are needed. The electroencephalographic (EEG) data have been utilized to study the differences of brain signals between alcoholics and healthy controls that could further developed as an automatic screening tool for alcoholics. In this work, resting-state EEG-derived features were utilized as input data to the proposed feature selection and classification method. The aim was to perform automatic classification of AUD patients and healthy controls. The validation of the proposed method involved real-EEG data acquired from 30 AUD patients and 30 age-matched healthy controls. The resting-state EEG-derived features such as synchronization likelihood (SL) were computed involving 19 scalp locations resulted into 513 features. Furthermore, the features were rank-ordered to select the most discriminant features involving a rank-based feature selection method according to a criterion, i.e., receiver operating characteristics (ROC). Consequently, a reduced set of most discriminant features was identified and utilized further during classification of AUD patients and healthy controls. In this study, three different classification models such as Support Vector Machine (SVM), Naïve Bayesian (NB), and Logistic Regression (LR) were used. The study resulted into SVM classification accuracy=98%, sensitivity=99.9%, specificity=95%, and f-measure=0.97; LR classification accuracy=91.7%, sensitivity=86.66%, specificity=96.6%, and f-measure=0.90; NB classification accuracy=93.6%, sensitivity=100%, specificity=87.9%, and f-measure=0.95. The SL features could be utilized as objective markers to screen the AUD patients and healthy controls. Copyright © 2017 Elsevier B.V. All rights reserved.

Predicting Response to Neoadjuvant Chemoradiotherapy in Esophageal Cancer with Textural Features Derived from Pretreatment 18F-FDG PET/CT Imaging.

PubMed

Beukinga, Roelof J; Hulshoff, Jan B; van Dijk, Lisanne V; Muijs, Christina T; Burgerhof, Johannes G M; Kats-Ugurlu, Gursah; Slart, Riemer H J A; Slump, Cornelis H; Mul, Véronique E M; Plukker, John Th M

2017-05-01

Adequate prediction of tumor response to neoadjuvant chemoradiotherapy (nCRT) in esophageal cancer (EC) patients is important in a more personalized treatment. The current best clinical method to predict pathologic complete response is SUV max in 18 F-FDG PET/CT imaging. To improve the prediction of response, we constructed a model to predict complete response to nCRT in EC based on pretreatment clinical parameters and 18 F-FDG PET/CT-derived textural features. Methods: From a prospectively maintained single-institution database, we reviewed 97 consecutive patients with locally advanced EC and a pretreatment 18 F-FDG PET/CT scan between 2009 and 2015. All patients were treated with nCRT (carboplatin/paclitaxel/41.4 Gy) followed by esophagectomy. We analyzed clinical, geometric, and pretreatment textural features extracted from both 18 F-FDG PET and CT. The current most accurate prediction model with SUV max as a predictor variable was compared with 6 different response prediction models constructed using least absolute shrinkage and selection operator regularized logistic regression. Internal validation was performed to estimate the model's performances. Pathologic response was defined as complete versus incomplete response (Mandard tumor regression grade system 1 vs. 2-5). Results: Pathologic examination revealed 19 (19.6%) complete and 78 (80.4%) incomplete responders. Least absolute shrinkage and selection operator regularization selected the clinical parameters: histologic type and clinical T stage, the 18 F-FDG PET-derived textural feature long run low gray level emphasis, and the CT-derived textural feature run percentage. Introducing these variables to a logistic regression analysis showed areas under the receiver-operating-characteristic curve (AUCs) of 0.78 compared with 0.58 in the SUV max model. The discrimination slopes were 0.17 compared with 0.01, respectively. After internal validation, the AUCs decreased to 0.74 and 0.54, respectively. Conclusion: The predictive values of the constructed models were superior to the standard method (SUV max ). These results can be considered as an initial step in predicting tumor response to nCRT in locally advanced EC. Further research in refining the predictive value of these models is needed to justify omission of surgery. © 2017 by the Society of Nuclear Medicine and Molecular Imaging.

Train axle bearing fault detection using a feature selection scheme based multi-scale morphological filter

NASA Astrophysics Data System (ADS)

Li, Yifan; Liang, Xihui; Lin, Jianhui; Chen, Yuejian; Liu, Jianxin

2018-02-01

This paper presents a novel signal processing scheme, feature selection based multi-scale morphological filter (MMF), for train axle bearing fault detection. In this scheme, more than 30 feature indicators of vibration signals are calculated for axle bearings with different conditions and the features which can reflect fault characteristics more effectively and representatively are selected using the max-relevance and min-redundancy principle. Then, a filtering scale selection approach for MMF based on feature selection and grey relational analysis is proposed. The feature selection based MMF method is tested on diagnosis of artificially created damages of rolling bearings of railway trains. Experimental results show that the proposed method has a superior performance in extracting fault features of defective train axle bearings. In addition, comparisons are performed with the kurtosis criterion based MMF and the spectral kurtosis criterion based MMF. The proposed feature selection based MMF method outperforms these two methods in detection of train axle bearing faults.

Web document ranking via active learning and kernel principal component analysis

NASA Astrophysics Data System (ADS)

Cai, Fei; Chen, Honghui; Shu, Zhen

2015-09-01

Web document ranking arises in many information retrieval (IR) applications, such as the search engine, recommendation system and online advertising. A challenging issue is how to select the representative query-document pairs and informative features as well for better learning and exploring new ranking models to produce an acceptable ranking list of candidate documents of each query. In this study, we propose an active sampling (AS) plus kernel principal component analysis (KPCA) based ranking model, viz. AS-KPCA Regression, to study the document ranking for a retrieval system, i.e. how to choose the representative query-document pairs and features for learning. More precisely, we fill those documents gradually into the training set by AS such that each of which will incur the highest expected DCG loss if unselected. Then, the KPCA is performed via projecting the selected query-document pairs onto p-principal components in the feature space to complete the regression. Hence, we can cut down the computational overhead and depress the impact incurred by noise simultaneously. To the best of our knowledge, we are the first to perform the document ranking via dimension reductions in two dimensions, namely, the number of documents and features simultaneously. Our experiments demonstrate that the performance of our approach is better than that of the baseline methods on the public LETOR 4.0 datasets. Our approach brings an improvement against RankBoost as well as other baselines near 20% in terms of MAP metric and less improvements using P@K and NDCG@K, respectively. Moreover, our approach is particularly suitable for document ranking on the noisy dataset in practice.

Data mining and visualization of average images in a digital hand atlas

NASA Astrophysics Data System (ADS)

Zhang, Aifeng; Gertych, Arkadiusz; Liu, Brent J.; Huang, H. K.

2005-04-01

We have collected a digital hand atlas containing digitized left hand radiographs of normally developed children grouped accordingly by age, sex, and race. A set of features stored in a database reflecting patient's stage of skeletal development has been calculated by automatic image processing procedures. This paper addresses a new concept, "average" image in the digital hand atlas. The "average" reference image in the digital atlas is selected for each of the groups of normal developed children with the best representative skeletal maturity based on bony features. A data mining procedure was designed and applied to find the average image through average feature vector matching. It also provides a temporary solution for the missing feature problem through polynomial regression. As more cases are added to the digital hand atlas, it can grow to provide clinicians accurate reference images to aid the bone age assessment process.

Fourier power, subjective distance, and object categories all provide plausible models of BOLD responses in scene-selective visual areas

PubMed Central

Lescroart, Mark D.; Stansbury, Dustin E.; Gallant, Jack L.

2015-01-01

Perception of natural visual scenes activates several functional areas in the human brain, including the Parahippocampal Place Area (PPA), Retrosplenial Complex (RSC), and the Occipital Place Area (OPA). It is currently unclear what specific scene-related features are represented in these areas. Previous studies have suggested that PPA, RSC, and/or OPA might represent at least three qualitatively different classes of features: (1) 2D features related to Fourier power; (2) 3D spatial features such as the distance to objects in a scene; or (3) abstract features such as the categories of objects in a scene. To determine which of these hypotheses best describes the visual representation in scene-selective areas, we applied voxel-wise modeling (VM) to BOLD fMRI responses elicited by a set of 1386 images of natural scenes. VM provides an efficient method for testing competing hypotheses by comparing predictions of brain activity based on encoding models that instantiate each hypothesis. Here we evaluated three different encoding models that instantiate each of the three hypotheses listed above. We used linear regression to fit each encoding model to the fMRI data recorded from each voxel, and we evaluated each fit model by estimating the amount of variance it predicted in a withheld portion of the data set. We found that voxel-wise models based on Fourier power or the subjective distance to objects in each scene predicted much of the variance predicted by a model based on object categories. Furthermore, the response variance explained by these three models is largely shared, and the individual models explain little unique variance in responses. Based on an evaluation of previous studies and the data we present here, we conclude that there is currently no good basis to favor any one of the three alternative hypotheses about visual representation in scene-selective areas. We offer suggestions for further studies that may help resolve this issue. PMID:26594164

Identifying predictive features in drug response using machine learning: opportunities and challenges.

PubMed

Vidyasagar, Mathukumalli

2015-01-01

This article reviews several techniques from machine learning that can be used to study the problem of identifying a small number of features, from among tens of thousands of measured features, that can accurately predict a drug response. Prediction problems are divided into two categories: sparse classification and sparse regression. In classification, the clinical parameter to be predicted is binary, whereas in regression, the parameter is a real number. Well-known methods for both classes of problems are briefly discussed. These include the SVM (support vector machine) for classification and various algorithms such as ridge regression, LASSO (least absolute shrinkage and selection operator), and EN (elastic net) for regression. In addition, several well-established methods that do not directly fall into machine learning theory are also reviewed, including neural networks, PAM (pattern analysis for microarrays), SAM (significance analysis for microarrays), GSEA (gene set enrichment analysis), and k-means clustering. Several references indicative of the application of these methods to cancer biology are discussed.

Time-Gated Raman Spectroscopy for Quantitative Determination of Solid-State Forms of Fluorescent Pharmaceuticals.

PubMed

Lipiäinen, Tiina; Pessi, Jenni; Movahedi, Parisa; Koivistoinen, Juha; Kurki, Lauri; Tenhunen, Mari; Yliruusi, Jouko; Juppo, Anne M; Heikkonen, Jukka; Pahikkala, Tapio; Strachan, Clare J

2018-04-03

Raman spectroscopy is widely used for quantitative pharmaceutical analysis, but a common obstacle to its use is sample fluorescence masking the Raman signal. Time-gating provides an instrument-based method for rejecting fluorescence through temporal resolution of the spectral signal and allows Raman spectra of fluorescent materials to be obtained. An additional practical advantage is that analysis is possible in ambient lighting. This study assesses the efficacy of time-gated Raman spectroscopy for the quantitative measurement of fluorescent pharmaceuticals. Time-gated Raman spectroscopy with a 128 × (2) × 4 CMOS SPAD detector was applied for quantitative analysis of ternary mixtures of solid-state forms of the model drug, piroxicam (PRX). Partial least-squares (PLS) regression allowed quantification, with Raman-active time domain selection (based on visual inspection) improving performance. Model performance was further improved by using kernel-based regularized least-squares (RLS) regression with greedy feature selection in which the data use in both the Raman shift and time dimensions was statistically optimized. Overall, time-gated Raman spectroscopy, especially with optimized data analysis in both the spectral and time dimensions, shows potential for sensitive and relatively routine quantitative analysis of photoluminescent pharmaceuticals during drug development and manufacturing.

Quantum-enhanced feature selection with forward selection and backward elimination

NASA Astrophysics Data System (ADS)

He, Zhimin; Li, Lvzhou; Huang, Zhiming; Situ, Haozhen

2018-07-01

Feature selection is a well-known preprocessing technique in machine learning, which can remove irrelevant features to improve the generalization capability of a classifier and reduce training and inference time. However, feature selection is time-consuming, particularly for the applications those have thousands of features, such as image retrieval, text mining and microarray data analysis. It is crucial to accelerate the feature selection process. We propose a quantum version of wrapper-based feature selection, which converts a classical feature selection to its quantum counterpart. It is valuable for machine learning on quantum computer. In this paper, we focus on two popular kinds of feature selection methods, i.e., wrapper-based forward selection and backward elimination. The proposed feature selection algorithm can quadratically accelerate the classical one.

Prediction of near-term breast cancer risk using local region-based bilateral asymmetry features in mammography

NASA Astrophysics Data System (ADS)

Li, Yane; Fan, Ming; Li, Lihua; Zheng, Bin

2017-03-01

This study proposed a near-term breast cancer risk assessment model based on local region bilateral asymmetry features in Mammography. The database includes 566 cases who underwent at least two sequential FFDM examinations. The `prior' examination in the two series all interpreted as negative (not recalled). In the "current" examination, 283 women were diagnosed cancers and 283 remained negative. Age of cancers and negative cases completely matched. These cases were divided into three subgroups according to age: 152 cases among the 37-49 age-bracket, 220 cases in the age-bracket 50- 60, and 194 cases with the 61-86 age-bracket. For each image, two local regions including strip-based regions and difference-of-Gaussian basic element regions were segmented. After that, structural variation features among pixel values and structural similarity features were computed for strip regions. Meanwhile, positional features were extracted for basic element regions. The absolute subtraction value was computed between each feature of the left and right local-regions. Next, a multi-layer perception classifier was implemented to assess performance of features for prediction. Features were then selected according stepwise regression analysis. The AUC achieved 0.72, 0.75 and 0.71 for these 3 age-based subgroups, respectively. The maximum adjustable odds ratios were 12.4, 20.56 and 4.91 for these three groups, respectively. This study demonstrate that the local region-based bilateral asymmetry features extracted from CC-view mammography could provide useful information to predict near-term breast cancer risk.

Spectral Regression Discriminant Analysis for Hyperspectral Image Classification

NASA Astrophysics Data System (ADS)

Pan, Y.; Wu, J.; Huang, H.; Liu, J.

2012-08-01

Dimensionality reduction algorithms, which aim to select a small set of efficient and discriminant features, have attracted great attention for Hyperspectral Image Classification. The manifold learning methods are popular for dimensionality reduction, such as Locally Linear Embedding, Isomap, and Laplacian Eigenmap. However, a disadvantage of many manifold learning methods is that their computations usually involve eigen-decomposition of dense matrices which is expensive in both time and memory. In this paper, we introduce a new dimensionality reduction method, called Spectral Regression Discriminant Analysis (SRDA). SRDA casts the problem of learning an embedding function into a regression framework, which avoids eigen-decomposition of dense matrices. Also, with the regression based framework, different kinds of regularizes can be naturally incorporated into our algorithm which makes it more flexible. It can make efficient use of data points to discover the intrinsic discriminant structure in the data. Experimental results on Washington DC Mall and AVIRIS Indian Pines hyperspectral data sets demonstrate the effectiveness of the proposed method.

Genetic Particle Swarm Optimization-Based Feature Selection for Very-High-Resolution Remotely Sensed Imagery Object Change Detection.

PubMed

Chen, Qiang; Chen, Yunhao; Jiang, Weiguo

2016-07-30

In the field of multiple features Object-Based Change Detection (OBCD) for very-high-resolution remotely sensed images, image objects have abundant features and feature selection affects the precision and efficiency of OBCD. Through object-based image analysis, this paper proposes a Genetic Particle Swarm Optimization (GPSO)-based feature selection algorithm to solve the optimization problem of feature selection in multiple features OBCD. We select the Ratio of Mean to Variance (RMV) as the fitness function of GPSO, and apply the proposed algorithm to the object-based hybrid multivariate alternative detection model. Two experiment cases on Worldview-2/3 images confirm that GPSO can significantly improve the speed of convergence, and effectively avoid the problem of premature convergence, relative to other feature selection algorithms. According to the accuracy evaluation of OBCD, GPSO is superior at overall accuracy (84.17% and 83.59%) and Kappa coefficient (0.6771 and 0.6314) than other algorithms. Moreover, the sensitivity analysis results show that the proposed algorithm is not easily influenced by the initial parameters, but the number of features to be selected and the size of the particle swarm would affect the algorithm. The comparison experiment results reveal that RMV is more suitable than other functions as the fitness function of GPSO-based feature selection algorithm.

Predicting volume of distribution with decision tree-based regression methods using predicted tissue:plasma partition coefficients.

PubMed

Freitas, Alex A; Limbu, Kriti; Ghafourian, Taravat

2015-01-01

Volume of distribution is an important pharmacokinetic property that indicates the extent of a drug's distribution in the body tissues. This paper addresses the problem of how to estimate the apparent volume of distribution at steady state (Vss) of chemical compounds in the human body using decision tree-based regression methods from the area of data mining (or machine learning). Hence, the pros and cons of several different types of decision tree-based regression methods have been discussed. The regression methods predict Vss using, as predictive features, both the compounds' molecular descriptors and the compounds' tissue:plasma partition coefficients (Kt:p) - often used in physiologically-based pharmacokinetics. Therefore, this work has assessed whether the data mining-based prediction of Vss can be made more accurate by using as input not only the compounds' molecular descriptors but also (a subset of) their predicted Kt:p values. Comparison of the models that used only molecular descriptors, in particular, the Bagging decision tree (mean fold error of 2.33), with those employing predicted Kt:p values in addition to the molecular descriptors, such as the Bagging decision tree using adipose Kt:p (mean fold error of 2.29), indicated that the use of predicted Kt:p values as descriptors may be beneficial for accurate prediction of Vss using decision trees if prior feature selection is applied. Decision tree based models presented in this work have an accuracy that is reasonable and similar to the accuracy of reported Vss inter-species extrapolations in the literature. The estimation of Vss for new compounds in drug discovery will benefit from methods that are able to integrate large and varied sources of data and flexible non-linear data mining methods such as decision trees, which can produce interpretable models. Graphical AbstractDecision trees for the prediction of tissue partition coefficient and volume of distribution of drugs.

Estimation of diffusion coefficients from voltammetric signals by support vector and gaussian process regression

PubMed Central

2014-01-01

Background Support vector regression (SVR) and Gaussian process regression (GPR) were used for the analysis of electroanalytical experimental data to estimate diffusion coefficients. Results For simulated cyclic voltammograms based on the EC, Eqr, and EqrC mechanisms these regression algorithms in combination with nonlinear kernel/covariance functions yielded diffusion coefficients with higher accuracy as compared to the standard approach of calculating diffusion coefficients relying on the Nicholson-Shain equation. The level of accuracy achieved by SVR and GPR is virtually independent of the rate constants governing the respective reaction steps. Further, the reduction of high-dimensional voltammetric signals by manual selection of typical voltammetric peak features decreased the performance of both regression algorithms compared to a reduction by downsampling or principal component analysis. After training on simulated data sets, diffusion coefficients were estimated by the regression algorithms for experimental data comprising voltammetric signals for three organometallic complexes. Conclusions Estimated diffusion coefficients closely matched the values determined by the parameter fitting method, but reduced the required computational time considerably for one of the reaction mechanisms. The automated processing of voltammograms according to the regression algorithms yields better results than the conventional analysis of peak-related data. PMID:24987463

Combination of radiological and gray level co-occurrence matrix textural features used to distinguish solitary pulmonary nodules by computed tomography.

PubMed

Wu, Haifeng; Sun, Tao; Wang, Jingjing; Li, Xia; Wang, Wei; Huo, Da; Lv, Pingxin; He, Wen; Wang, Keyang; Guo, Xiuhua

2013-08-01

The objective of this study was to investigate the method of the combination of radiological and textural features for the differentiation of malignant from benign solitary pulmonary nodules by computed tomography. Features including 13 gray level co-occurrence matrix textural features and 12 radiological features were extracted from 2,117 CT slices, which came from 202 (116 malignant and 86 benign) patients. Lasso-type regularization to a nonlinear regression model was applied to select predictive features and a BP artificial neural network was used to build the diagnostic model. Eight radiological and two textural features were obtained after the Lasso-type regularization procedure. Twelve radiological features alone could reach an area under the ROC curve (AUC) of 0.84 in differentiating between malignant and benign lesions. The 10 selected characters improved the AUC to 0.91. The evaluation results showed that the method of selecting radiological and textural features appears to yield more effective in the distinction of malignant from benign solitary pulmonary nodules by computed tomography.

Optimizing data collection for public health decisions: a data mining approach

PubMed Central

2014-01-01

Background Collecting data can be cumbersome and expensive. Lack of relevant, accurate and timely data for research to inform policy may negatively impact public health. The aim of this study was to test if the careful removal of items from two community nutrition surveys guided by a data mining technique called feature selection, can (a) identify a reduced dataset, while (b) not damaging the signal inside that data. Methods The Nutrition Environment Measures Surveys for stores (NEMS-S) and restaurants (NEMS-R) were completed on 885 retail food outlets in two counties in West Virginia between May and November of 2011. A reduced dataset was identified for each outlet type using feature selection. Coefficients from linear regression modeling were used to weight items in the reduced datasets. Weighted item values were summed with the error term to compute reduced item survey scores. Scores produced by the full survey were compared to the reduced item scores using a Wilcoxon rank-sum test. Results Feature selection identified 9 store and 16 restaurant survey items as significant predictors of the score produced from the full survey. The linear regression models built from the reduced feature sets had R2 values of 92% and 94% for restaurant and grocery store data, respectively. Conclusions While there are many potentially important variables in any domain, the most useful set may only be a small subset. The use of feature selection in the initial phase of data collection to identify the most influential variables may be a useful tool to greatly reduce the amount of data needed thereby reducing cost. PMID:24919484

Optimizing data collection for public health decisions: a data mining approach.

PubMed

Partington, Susan N; Papakroni, Vasil; Menzies, Tim

2014-06-12

Collecting data can be cumbersome and expensive. Lack of relevant, accurate and timely data for research to inform policy may negatively impact public health. The aim of this study was to test if the careful removal of items from two community nutrition surveys guided by a data mining technique called feature selection, can (a) identify a reduced dataset, while (b) not damaging the signal inside that data. The Nutrition Environment Measures Surveys for stores (NEMS-S) and restaurants (NEMS-R) were completed on 885 retail food outlets in two counties in West Virginia between May and November of 2011. A reduced dataset was identified for each outlet type using feature selection. Coefficients from linear regression modeling were used to weight items in the reduced datasets. Weighted item values were summed with the error term to compute reduced item survey scores. Scores produced by the full survey were compared to the reduced item scores using a Wilcoxon rank-sum test. Feature selection identified 9 store and 16 restaurant survey items as significant predictors of the score produced from the full survey. The linear regression models built from the reduced feature sets had R2 values of 92% and 94% for restaurant and grocery store data, respectively. While there are many potentially important variables in any domain, the most useful set may only be a small subset. The use of feature selection in the initial phase of data collection to identify the most influential variables may be a useful tool to greatly reduce the amount of data needed thereby reducing cost.

Improving stability of prediction models based on correlated omics data by using network approaches.

PubMed

Tissier, Renaud; Houwing-Duistermaat, Jeanine; Rodríguez-Girondo, Mar

2018-01-01

Building prediction models based on complex omics datasets such as transcriptomics, proteomics, metabolomics remains a challenge in bioinformatics and biostatistics. Regularized regression techniques are typically used to deal with the high dimensionality of these datasets. However, due to the presence of correlation in the datasets, it is difficult to select the best model and application of these methods yields unstable results. We propose a novel strategy for model selection where the obtained models also perform well in terms of overall predictability. Several three step approaches are considered, where the steps are 1) network construction, 2) clustering to empirically derive modules or pathways, and 3) building a prediction model incorporating the information on the modules. For the first step, we use weighted correlation networks and Gaussian graphical modelling. Identification of groups of features is performed by hierarchical clustering. The grouping information is included in the prediction model by using group-based variable selection or group-specific penalization. We compare the performance of our new approaches with standard regularized regression via simulations. Based on these results we provide recommendations for selecting a strategy for building a prediction model given the specific goal of the analysis and the sizes of the datasets. Finally we illustrate the advantages of our approach by application of the methodology to two problems, namely prediction of body mass index in the DIetary, Lifestyle, and Genetic determinants of Obesity and Metabolic syndrome study (DILGOM) and prediction of response of each breast cancer cell line to treatment with specific drugs using a breast cancer cell lines pharmacogenomics dataset.

Development and evaluation of an ambulatory stress monitor based on wearable sensors.

PubMed

Choi, Jongyoon; Ahmed, Beena; Gutierrez-Osuna, Ricardo

2012-03-01

Chronic stress is endemic to modern society. However, as it is unfeasible for physicians to continuously monitor stress levels, its diagnosis is nontrivial. Wireless body sensor networks offer opportunities to ubiquitously detect and monitor mental stress levels, enabling improved diagnosis, and early treatment. This article describes the development of a wearable sensor platform to monitor a number of physiological correlates of mental stress. We discuss tradeoffs in both system design and sensor selection to balance information content and wearability. Using experimental signals collected from the wearable sensor, we describe a selected number of physiological features that show good correlation with mental stress. In particular, we propose a new spectral feature that estimates the balance of the autonomic nervous system by combining information from the power spectral density of respiration and heart rate variability. We validate the effectiveness of our approach on a binary discrimination problem when subjects are placed under two psychophysiological conditions: mental stress and relaxation. When used in a logistic regression model, our feature set is able to discriminate between these two mental states with a success rate of 81% across subjects. © 2012 IEEE

Distributed Lag Models: Examining Associations between the Built Environment and Health

PubMed Central

Baek, Jonggyu; Sánchez, Brisa N.; Berrocal, Veronica J.; Sanchez-Vaznaugh, Emma V.

2016-01-01

Built environment factors constrain individual level behaviors and choices, and thus are receiving increasing attention to assess their influence on health. Traditional regression methods have been widely used to examine associations between built environment measures and health outcomes, where a fixed, pre-specified spatial scale (e.g., 1 mile buffer) is used to construct environment measures. However, the spatial scale for these associations remains largely unknown and misspecifying it introduces bias. We propose the use of distributed lag models (DLMs) to describe the association between built environment features and health as a function of distance from the locations of interest and circumvent a-priori selection of a spatial scale. Based on simulation studies, we demonstrate that traditional regression models produce associations biased away from the null when there is spatial correlation among the built environment features. Inference based on DLMs is robust under a range of scenarios of the built environment. We use this innovative application of DLMs to examine the association between the availability of convenience stores near California public schools, which may affect children’s dietary choices both through direct access to junk food and exposure to advertisement, and children’s body mass index z-scores (BMIz). PMID:26414942

Attentional Selection Can Be Predicted by Reinforcement Learning of Task-relevant Stimulus Features Weighted by Value-independent Stickiness.

PubMed

Balcarras, Matthew; Ardid, Salva; Kaping, Daniel; Everling, Stefan; Womelsdorf, Thilo

2016-02-01

Attention includes processes that evaluate stimuli relevance, select the most relevant stimulus against less relevant stimuli, and bias choice behavior toward the selected information. It is not clear how these processes interact. Here, we captured these processes in a reinforcement learning framework applied to a feature-based attention task that required macaques to learn and update the value of stimulus features while ignoring nonrelevant sensory features, locations, and action plans. We found that value-based reinforcement learning mechanisms could account for feature-based attentional selection and choice behavior but required a value-independent stickiness selection process to explain selection errors while at asymptotic behavior. By comparing different reinforcement learning schemes, we found that trial-by-trial selections were best predicted by a model that only represents expected values for the task-relevant feature dimension, with nonrelevant stimulus features and action plans having only a marginal influence on covert selections. These findings show that attentional control subprocesses can be described by (1) the reinforcement learning of feature values within a restricted feature space that excludes irrelevant feature dimensions, (2) a stochastic selection process on feature-specific value representations, and (3) value-independent stickiness toward previous feature selections akin to perseveration in the motor domain. We speculate that these three mechanisms are implemented by distinct but interacting brain circuits and that the proposed formal account of feature-based stimulus selection will be important to understand how attentional subprocesses are implemented in primate brain networks.

Formation enthalpies for transition metal alloys using machine learning

NASA Astrophysics Data System (ADS)

Ubaru, Shashanka; Miedlar, Agnieszka; Saad, Yousef; Chelikowsky, James R.

2017-06-01

The enthalpy of formation is an important thermodynamic property. Developing fast and accurate methods for its prediction is of practical interest in a variety of applications. Material informatics techniques based on machine learning have recently been introduced in the literature as an inexpensive means of exploiting materials data, and can be used to examine a variety of thermodynamics properties. We investigate the use of such machine learning tools for predicting the formation enthalpies of binary intermetallic compounds that contain at least one transition metal. We consider certain easily available properties of the constituting elements complemented by some basic properties of the compounds, to predict the formation enthalpies. We show how choosing these properties (input features) based on a literature study (using prior physics knowledge) seems to outperform machine learning based feature selection methods such as sensitivity analysis and LASSO (least absolute shrinkage and selection operator) based methods. A nonlinear kernel based support vector regression method is employed to perform the predictions. The predictive ability of our model is illustrated via several experiments on a dataset containing 648 binary alloys. We train and validate the model using the formation enthalpies calculated using a model by Miedema, which is a popular semiempirical model used for the prediction of formation enthalpies of metal alloys.

Structure-based predictions of 13C-NMR chemical shifts for a series of 2-functionalized 5-(methylsulfonyl)-1-phenyl-1H-indoles derivatives using GA-based MLR method

NASA Astrophysics Data System (ADS)

Ghavami, Raouf; Sadeghi, Faridoon; Rasouli, Zolikha; Djannati, Farhad

2012-12-01

Experimental values for the 13C NMR chemical shifts (ppm, TMS = 0) at 300 K ranging from 96.28 ppm (C4' of indole derivative 17) to 159.93 ppm (C4' of indole derivative 23) relative to deuteride chloroform (CDCl3, 77.0 ppm) or dimethylsulfoxide (DMSO, 39.50 ppm) as internal reference in CDCl3 or DMSO-d6 solutions have been collected from literature for thirty 2-functionalized 5-(methylsulfonyl)-1-phenyl-1H-indole derivatives containing different substituted groups. An effective quantitative structure-property relationship (QSPR) models were built using hybrid method combining genetic algorithm (GA) based on stepwise selection multiple linear regression (SWS-MLR) as feature-selection tools and correlation models between each carbon atom of indole derivative and calculated descriptors. Each compound was depicted by molecular structural descriptors that encode constitutional, topological, geometrical, electrostatic, and quantum chemical features. The accuracy of all developed models were confirmed using different types of internal and external procedures and various statistical tests. Furthermore, the domain of applicability for each model which indicates the area of reliable predictions was defined.

Genetic Particle Swarm Optimization–Based Feature Selection for Very-High-Resolution Remotely Sensed Imagery Object Change Detection

PubMed Central

Chen, Qiang; Chen, Yunhao; Jiang, Weiguo

2016-01-01

In the field of multiple features Object-Based Change Detection (OBCD) for very-high-resolution remotely sensed images, image objects have abundant features and feature selection affects the precision and efficiency of OBCD. Through object-based image analysis, this paper proposes a Genetic Particle Swarm Optimization (GPSO)-based feature selection algorithm to solve the optimization problem of feature selection in multiple features OBCD. We select the Ratio of Mean to Variance (RMV) as the fitness function of GPSO, and apply the proposed algorithm to the object-based hybrid multivariate alternative detection model. Two experiment cases on Worldview-2/3 images confirm that GPSO can significantly improve the speed of convergence, and effectively avoid the problem of premature convergence, relative to other feature selection algorithms. According to the accuracy evaluation of OBCD, GPSO is superior at overall accuracy (84.17% and 83.59%) and Kappa coefficient (0.6771 and 0.6314) than other algorithms. Moreover, the sensitivity analysis results show that the proposed algorithm is not easily influenced by the initial parameters, but the number of features to be selected and the size of the particle swarm would affect the algorithm. The comparison experiment results reveal that RMV is more suitable than other functions as the fitness function of GPSO-based feature selection algorithm. PMID:27483285

Local curvature analysis for classifying breast tumors: Preliminary analysis in dedicated breast CT

DOE Office of Scientific and Technical Information (OSTI.GOV)

Lee, Juhun, E-mail: leej15@upmc.edu; Nishikawa, Robert M.; Reiser, Ingrid

2015-09-15

Purpose: The purpose of this study is to measure the effectiveness of local curvature measures as novel image features for classifying breast tumors. Methods: A total of 119 breast lesions from 104 noncontrast dedicated breast computed tomography images of women were used in this study. Volumetric segmentation was done using a seed-based segmentation algorithm and then a triangulated surface was extracted from the resulting segmentation. Total, mean, and Gaussian curvatures were then computed. Normalized curvatures were used as classification features. In addition, traditional image features were also extracted and a forward feature selection scheme was used to select the optimalmore » feature set. Logistic regression was used as a classifier and leave-one-out cross-validation was utilized to evaluate the classification performances of the features. The area under the receiver operating characteristic curve (AUC, area under curve) was used as a figure of merit. Results: Among curvature measures, the normalized total curvature (C{sub T}) showed the best classification performance (AUC of 0.74), while the others showed no classification power individually. Five traditional image features (two shape, two margin, and one texture descriptors) were selected via the feature selection scheme and its resulting classifier achieved an AUC of 0.83. Among those five features, the radial gradient index (RGI), which is a margin descriptor, showed the best classification performance (AUC of 0.73). A classifier combining RGI and C{sub T} yielded an AUC of 0.81, which showed similar performance (i.e., no statistically significant difference) to the classifier with the above five traditional image features. Additional comparisons in AUC values between classifiers using different combinations of traditional image features and C{sub T} were conducted. The results showed that C{sub T} was able to replace the other four image features for the classification task. Conclusions: The normalized curvature measure contains useful information in classifying breast tumors. Using this, one can reduce the number of features in a classifier, which may result in more robust classifiers for different datasets.« less

ANALYSIS OF SAMPLING TECHNIQUES FOR IMBALANCED DATA: AN N=648 ADNI STUDY

PubMed Central

Dubey, Rashmi; Zhou, Jiayu; Wang, Yalin; Thompson, Paul M.; Ye, Jieping

2013-01-01

Many neuroimaging applications deal with imbalanced imaging data. For example, in Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset, the mild cognitive impairment (MCI) cases eligible for the study are nearly two times the Alzheimer’s disease (AD) patients for structural magnetic resonance imaging (MRI) modality and six times the control cases for proteomics modality. Constructing an accurate classifier from imbalanced data is a challenging task. Traditional classifiers that aim to maximize the overall prediction accuracy tend to classify all data into the majority class. In this paper, we study an ensemble system of feature selection and data sampling for the class imbalance problem. We systematically analyze various sampling techniques by examining the efficacy of different rates and types of undersampling, oversampling, and a combination of over and under sampling approaches. We thoroughly examine six widely used feature selection algorithms to identify significant biomarkers and thereby reduce the complexity of the data. The efficacy of the ensemble techniques is evaluated using two different classifiers including Random Forest and Support Vector Machines based on classification accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity measures. Our extensive experimental results show that for various problem settings in ADNI, (1). a balanced training set obtained with K-Medoids technique based undersampling gives the best overall performance among different data sampling techniques and no sampling approach; and (2). sparse logistic regression with stability selection achieves competitive performance among various feature selection algorithms. Comprehensive experiments with various settings show that our proposed ensemble model of multiple undersampled datasets yields stable and promising results. PMID:24176869

Analysis of sampling techniques for imbalanced data: An n = 648 ADNI study.

PubMed

Dubey, Rashmi; Zhou, Jiayu; Wang, Yalin; Thompson, Paul M; Ye, Jieping

2014-02-15

Many neuroimaging applications deal with imbalanced imaging data. For example, in Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, the mild cognitive impairment (MCI) cases eligible for the study are nearly two times the Alzheimer's disease (AD) patients for structural magnetic resonance imaging (MRI) modality and six times the control cases for proteomics modality. Constructing an accurate classifier from imbalanced data is a challenging task. Traditional classifiers that aim to maximize the overall prediction accuracy tend to classify all data into the majority class. In this paper, we study an ensemble system of feature selection and data sampling for the class imbalance problem. We systematically analyze various sampling techniques by examining the efficacy of different rates and types of undersampling, oversampling, and a combination of over and undersampling approaches. We thoroughly examine six widely used feature selection algorithms to identify significant biomarkers and thereby reduce the complexity of the data. The efficacy of the ensemble techniques is evaluated using two different classifiers including Random Forest and Support Vector Machines based on classification accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity measures. Our extensive experimental results show that for various problem settings in ADNI, (1) a balanced training set obtained with K-Medoids technique based undersampling gives the best overall performance among different data sampling techniques and no sampling approach; and (2) sparse logistic regression with stability selection achieves competitive performance among various feature selection algorithms. Comprehensive experiments with various settings show that our proposed ensemble model of multiple undersampled datasets yields stable and promising results. © 2013 Elsevier Inc. All rights reserved.

Sparse representation of multi parametric DCE-MRI features using K-SVD for classifying gene expression based breast cancer recurrence risk

NASA Astrophysics Data System (ADS)

Mahrooghy, Majid; Ashraf, Ahmed B.; Daye, Dania; Mies, Carolyn; Rosen, Mark; Feldman, Michael; Kontos, Despina

2014-03-01

We evaluate the prognostic value of sparse representation-based features by applying the K-SVD algorithm on multiparametric kinetic, textural, and morphologic features in breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). K-SVD is an iterative dimensionality reduction method that optimally reduces the initial feature space by updating the dictionary columns jointly with the sparse representation coefficients. Therefore, by using K-SVD, we not only provide sparse representation of the features and condense the information in a few coefficients but also we reduce the dimensionality. The extracted K-SVD features are evaluated by a machine learning algorithm including a logistic regression classifier for the task of classifying high versus low breast cancer recurrence risk as determined by a validated gene expression assay. The features are evaluated using ROC curve analysis and leave one-out cross validation for different sparse representation and dimensionality reduction numbers. Optimal sparse representation is obtained when the number of dictionary elements is 4 (K=4) and maximum non-zero coefficients is 2 (L=2). We compare K-SVD with ANOVA based feature selection for the same prognostic features. The ROC results show that the AUC of the K-SVD based (K=4, L=2), the ANOVA based, and the original features (i.e., no dimensionality reduction) are 0.78, 0.71. and 0.68, respectively. From the results, it can be inferred that by using sparse representation of the originally extracted multi-parametric, high-dimensional data, we can condense the information on a few coefficients with the highest predictive value. In addition, the dimensionality reduction introduced by K-SVD can prevent models from over-fitting.

Accurate Diabetes Risk Stratification Using Machine Learning: Role of Missing Value and Outliers.

PubMed

Maniruzzaman, Md; Rahman, Md Jahanur; Al-MehediHasan, Md; Suri, Harman S; Abedin, Md Menhazul; El-Baz, Ayman; Suri, Jasjit S

2018-04-10

Diabetes mellitus is a group of metabolic diseases in which blood sugar levels are too high. About 8.8% of the world was diabetic in 2017. It is projected that this will reach nearly 10% by 2045. The major challenge is that when machine learning-based classifiers are applied to such data sets for risk stratification, leads to lower performance. Thus, our objective is to develop an optimized and robust machine learning (ML) system under the assumption that missing values or outliers if replaced by a median configuration will yield higher risk stratification accuracy. This ML-based risk stratification is designed, optimized and evaluated, where: (i) the features are extracted and optimized from the six feature selection techniques (random forest, logistic regression, mutual information, principal component analysis, analysis of variance, and Fisher discriminant ratio) and combined with ten different types of classifiers (linear discriminant analysis, quadratic discriminant analysis, naïve Bayes, Gaussian process classification, support vector machine, artificial neural network, Adaboost, logistic regression, decision tree, and random forest) under the hypothesis that both missing values and outliers when replaced by computed medians will improve the risk stratification accuracy. Pima Indian diabetic dataset (768 patients: 268 diabetic and 500 controls) was used. Our results demonstrate that on replacing the missing values and outliers by group median and median values, respectively and further using the combination of random forest feature selection and random forest classification technique yields an accuracy, sensitivity, specificity, positive predictive value, negative predictive value and area under the curve as: 92.26%, 95.96%, 79.72%, 91.14%, 91.20%, and 0.93, respectively. This is an improvement of 10% over previously developed techniques published in literature. The system was validated for its stability and reliability. RF-based model showed the best performance when outliers are replaced by median values.

A Genetic Algorithm Based Support Vector Machine Model for Blood-Brain Barrier Penetration Prediction

PubMed Central

Zhang, Daqing; Xiao, Jianfeng; Zhou, Nannan; Luo, Xiaomin; Jiang, Hualiang; Chen, Kaixian

2015-01-01

Blood-brain barrier (BBB) is a highly complex physical barrier determining what substances are allowed to enter the brain. Support vector machine (SVM) is a kernel-based machine learning method that is widely used in QSAR study. For a successful SVM model, the kernel parameters for SVM and feature subset selection are the most important factors affecting prediction accuracy. In most studies, they are treated as two independent problems, but it has been proven that they could affect each other. We designed and implemented genetic algorithm (GA) to optimize kernel parameters and feature subset selection for SVM regression and applied it to the BBB penetration prediction. The results show that our GA/SVM model is more accurate than other currently available log BB models. Therefore, to optimize both SVM parameters and feature subset simultaneously with genetic algorithm is a better approach than other methods that treat the two problems separately. Analysis of our log BB model suggests that carboxylic acid group, polar surface area (PSA)/hydrogen-bonding ability, lipophilicity, and molecular charge play important role in BBB penetration. Among those properties relevant to BBB penetration, lipophilicity could enhance the BBB penetration while all the others are negatively correlated with BBB penetration. PMID:26504797

Visual attention based bag-of-words model for image classification

NASA Astrophysics Data System (ADS)

Wang, Qiwei; Wan, Shouhong; Yue, Lihua; Wang, Che

2014-04-01

Bag-of-words is a classical method for image classification. The core problem is how to count the frequency of the visual words and what visual words to select. In this paper, we propose a visual attention based bag-of-words model (VABOW model) for image classification task. The VABOW model utilizes visual attention method to generate a saliency map, and uses the saliency map as a weighted matrix to instruct the statistic process for the frequency of the visual words. On the other hand, the VABOW model combines shape, color and texture cues and uses L1 regularization logistic regression method to select the most relevant and most efficient features. We compare our approach with traditional bag-of-words based method on two datasets, and the result shows that our VABOW model outperforms the state-of-the-art method for image classification.

Integrating dimension reduction and out-of-sample extension in automated classification of ex vivo human patellar cartilage on phase contrast X-ray computed tomography.

PubMed

Nagarajan, Mahesh B; Coan, Paola; Huber, Markus B; Diemoz, Paul C; Wismüller, Axel

2015-01-01

Phase contrast X-ray computed tomography (PCI-CT) has been demonstrated as a novel imaging technique that can visualize human cartilage with high spatial resolution and soft tissue contrast. Different textural approaches have been previously investigated for characterizing chondrocyte organization on PCI-CT to enable classification of healthy and osteoarthritic cartilage. However, the large size of feature sets extracted in such studies motivates an investigation into algorithmic feature reduction for computing efficient feature representations without compromising their discriminatory power. For this purpose, geometrical feature sets derived from the scaling index method (SIM) were extracted from 1392 volumes of interest (VOI) annotated on PCI-CT images of ex vivo human patellar cartilage specimens. The extracted feature sets were subject to linear and non-linear dimension reduction techniques as well as feature selection based on evaluation of mutual information criteria. The reduced feature set was subsequently used in a machine learning task with support vector regression to classify VOIs as healthy or osteoarthritic; classification performance was evaluated using the area under the receiver-operating characteristic (ROC) curve (AUC). Our results show that the classification performance achieved by 9-D SIM-derived geometric feature sets (AUC: 0.96 ± 0.02) can be maintained with 2-D representations computed from both dimension reduction and feature selection (AUC values as high as 0.97 ± 0.02). Thus, such feature reduction techniques can offer a high degree of compaction to large feature sets extracted from PCI-CT images while maintaining their ability to characterize the underlying chondrocyte patterns.

Feature Selection, Flaring Size and Time-to-Flare Prediction Using Support Vector Regression, and Automated Prediction of Flaring Behavior Based on Spatio-Temporal Measures Using Hidden Markov Models

NASA Astrophysics Data System (ADS)

Al-Ghraibah, Amani

Solar flares release stored magnetic energy in the form of radiation and can have significant detrimental effects on earth including damage to technological infrastructure. Recent work has considered methods to predict future flare activity on the basis of quantitative measures of the solar magnetic field. Accurate advanced warning of solar flare occurrence is an area of increasing concern and much research is ongoing in this area. Our previous work 111] utilized standard pattern recognition and classification techniques to determine (classify) whether a region is expected to flare within a predictive time window, using a Relevance Vector Machine (RVM) classification method. We extracted 38 features which describing the complexity of the photospheric magnetic field, the result classification metrics will provide the baseline against which we compare our new work. We find a true positive rate (TPR) of 0.8, true negative rate (TNR) of 0.7, and true skill score (TSS) of 0.49. This dissertation proposes three basic topics; the first topic is an extension to our previous work [111, where we consider a feature selection method to determine an appropriate feature subset with cross validation classification based on a histogram analysis of selected features. Classification using the top five features resulting from this analysis yield better classification accuracies across a large unbalanced dataset. In particular, the feature subsets provide better discrimination of the many regions that flare where we find a TPR of 0.85, a TNR of 0.65 sightly lower than our previous work, and a TSS of 0.5 which has an improvement comparing with our previous work. In the second topic, we study the prediction of solar flare size and time-to-flare using support vector regression (SVR). When we consider flaring regions only, we find an average error in estimating flare size of approximately half a GOES class. When we additionally consider non-flaring regions, we find an increased average error of approximately 3/4 a GOES class. We also consider thresholding the regressed flare size for the experiment containing both flaring and non-flaring regions and find a TPR. of 0.69 and a TNR of 0.86 for flare prediction, consistent with our previous studies of flare prediction using the same magnetic complexity features. The results for both of these size regression experiments are consistent across a wide range of predictive time windows, indicating that the magnetic complexity features may be persistent in appearance long before flare activity. This conjecture is supported by our larger error rates of some 40 hours in the time-to-flare regression problem. The magnetic complexity features considered here appear to have discriminative potential for flare size, but their persistence in time makes them less discriminative for the time-to-flare problem. We also study the prediction of solar flare size and time-to-flare using two temporal features, namely the ▵- and ▵-▵-features, the same average size and time-to-flare regression error are found when these temporal features are used in size and time-to-flare prediction. In the third topic, we study the temporal evolution of active region magnetic fields using Hidden Markov Models (HMMs) which is one of the efficient temporal analyses found in literature. We extracted 38 features which describing the complexity of the photospheric magnetic field. These features are converted into a sequence of symbols using k-nearest neighbor search method. We study many parameters before prediction; like the length of the training window Wtrain which denotes to the number of history images use to train the flare and non-flare HMMs, and number of hidden states Q. In training phase, the model parameters of the HMM of each category are optimized so as to best describe the training symbol sequences. In testing phase, we use the best flare and non-flare models to predict/classify active regions as a flaring or non-flaring region using a sliding window method. The best prediction result is found where the length of the history training images are 15 images (i.e., Wtrain= 15) and the length of the sliding testing window is less than or equal to W train, the best result give a TPR of 0.79 consistent with previous flare prediction work, TNR of 0.87 arid TSS of 0.66, where both are higher than our previous flare prediction work. We find that the best number of hidden states which can describe the temporal evolution of the solar ARs is equal to five states, at the same time, a close resultant metrics are found using different number of states.

Multiple-output support vector machine regression with feature selection for arousal/valence space emotion assessment.

PubMed

Torres-Valencia, Cristian A; Álvarez, Mauricio A; Orozco-Gutiérrez, Alvaro A

2014-01-01

Human emotion recognition (HER) allows the assessment of an affective state of a subject. Until recently, such emotional states were described in terms of discrete emotions, like happiness or contempt. In order to cover a high range of emotions, researchers in the field have introduced different dimensional spaces for emotion description that allow the characterization of affective states in terms of several variables or dimensions that measure distinct aspects of the emotion. One of the most common of such dimensional spaces is the bidimensional Arousal/Valence space. To the best of our knowledge, all HER systems so far have modelled independently, the dimensions in these dimensional spaces. In this paper, we study the effect of modelling the output dimensions simultaneously and show experimentally the advantages in modeling them in this way. We consider a multimodal approach by including features from the Electroencephalogram and a few physiological signals. For modelling the multiple outputs, we employ a multiple output regressor based on support vector machines. We also include an stage of feature selection that is developed within an embedded approach known as Recursive Feature Elimination (RFE), proposed initially for SVM. The results show that several features can be eliminated using the multiple output support vector regressor with RFE without affecting the performance of the regressor. From the analysis of the features selected in smaller subsets via RFE, it can be observed that the signals that are more informative into the arousal and valence space discrimination are the EEG, Electrooculogram/Electromiogram (EOG/EMG) and the Galvanic Skin Response (GSR).

Variable importance in nonlinear kernels (VINK): classification of digitized histopathology.

PubMed

Ginsburg, Shoshana; Ali, Sahirzeeshan; Lee, George; Basavanhally, Ajay; Madabhushi, Anant

2013-01-01

Quantitative histomorphometry is the process of modeling appearance of disease morphology on digitized histopathology images via image-based features (e.g., texture, graphs). Due to the curse of dimensionality, building classifiers with large numbers of features requires feature selection (which may require a large training set) or dimensionality reduction (DR). DR methods map the original high-dimensional features in terms of eigenvectors and eigenvalues, which limits the potential for feature transparency or interpretability. Although methods exist for variable selection and ranking on embeddings obtained via linear DR schemes (e.g., principal components analysis (PCA)), similar methods do not yet exist for nonlinear DR (NLDR) methods. In this work we present a simple yet elegant method for approximating the mapping between the data in the original feature space and the transformed data in the kernel PCA (KPCA) embedding space; this mapping provides the basis for quantification of variable importance in nonlinear kernels (VINK). We show how VINK can be implemented in conjunction with the popular Isomap and Laplacian eigenmap algorithms. VINK is evaluated in the contexts of three different problems in digital pathology: (1) predicting five year PSA failure following radical prostatectomy, (2) predicting Oncotype DX recurrence risk scores for ER+ breast cancers, and (3) distinguishing good and poor outcome p16+ oropharyngeal tumors. We demonstrate that subsets of features identified by VINK provide similar or better classification or regression performance compared to the original high dimensional feature sets.

Association between background parenchymal enhancement of breast MRI and BIRADS rating change in the subsequent screening

NASA Astrophysics Data System (ADS)

Aghaei, Faranak; Mirniaharikandehei, Seyedehnafiseh; Hollingsworth, Alan B.; Stoug, Rebecca G.; Pearce, Melanie; Liu, Hong; Zheng, Bin

2018-03-01

Although breast magnetic resonance imaging (MRI) has been used as a breast cancer screening modality for high-risk women, its cancer detection yield remains low (i.e., <= 3%). Thus, increasing breast MRI screening efficacy and cancer detection yield is an important clinical issue in breast cancer screening. In this study, we investigated association between the background parenchymal enhancement (BPE) of breast MRI and the change of diagnostic (BIRADS) status in the next subsequent breast MRI screening. A dataset with 65 breast MRI screening cases was retrospectively assembled. All cases were rated BIRADS-2 (benign findings). In the subsequent screening, 4 cases were malignant (BIRADS-6), 48 remained BIRADS-2 and 13 were downgraded to negative (BIRADS-1). A computer-aided detection scheme was applied to process images of the first set of breast MRI screening. Total of 33 features were computed including texture feature and global BPE features. Texture features were computed from either a gray-level co-occurrence matrix or a gray level run length matrix. Ten global BPE features were also initially computed from two breast regions and bilateral difference between the left and right breasts. Box-plot based analysis shows positive association between texture features and BIRADS rating levels in the second screening. Furthermore, a logistic regression model was built using optimal features selected by a CFS based feature selection method. Using a leave-one-case-out based cross-validation method, classification yielded an overall 75% accuracy in predicting the improvement (or downgrade) of diagnostic status (to BIRAD-1) in the subsequent breast MRI screening. This study demonstrated potential of developing a new quantitative imaging marker to predict diagnostic status change in the short-term, which may help eliminate a high fraction of unnecessary repeated breast MRI screenings and increase the cancer detection yield.

Detecting nonsense for Chinese comments based on logistic regression

NASA Astrophysics Data System (ADS)

Zhuolin, Ren; Guang, Chen; Shu, Chen

2016-07-01

To understand cyber citizens' opinion accurately from Chinese news comments, the clear definition on nonsense is present, and a detection model based on logistic regression (LR) is proposed. The detection of nonsense can be treated as a binary-classification problem. Besides of traditional lexical features, we propose three kinds of features in terms of emotion, structure and relevance. By these features, we train an LR model and demonstrate its effect in understanding Chinese news comments. We find that each of proposed features can significantly promote the result. In our experiments, we achieve a prediction accuracy of 84.3% which improves the baseline 77.3% by 7%.

The fate of task-irrelevant visual motion: perceptual load versus feature-based attention.

PubMed

Taya, Shuichiro; Adams, Wendy J; Graf, Erich W; Lavie, Nilli

2009-11-18

We tested contrasting predictions derived from perceptual load theory and from recent feature-based selection accounts. Observers viewed moving, colored stimuli and performed low or high load tasks associated with one stimulus feature, either color or motion. The resultant motion aftereffect (MAE) was used to evaluate attentional allocation. We found that task-irrelevant visual features received less attention than co-localized task-relevant features of the same objects. Moreover, when color and motion features were co-localized yet perceived to belong to two distinct surfaces, feature-based selection was further increased at the expense of object-based co-selection. Load theory predicts that the MAE for task-irrelevant motion would be reduced with a higher load color task. However, this was not seen for co-localized features; perceptual load only modulated the MAE for task-irrelevant motion when this was spatially separated from the attended color location. Our results suggest that perceptual load effects are mediated by spatial selection and do not generalize to the feature domain. Feature-based selection operates to suppress processing of task-irrelevant, co-localized features, irrespective of perceptual load.

Modified CC-LR algorithm with three diverse feature sets for motor imagery tasks classification in EEG based brain-computer interface.

PubMed

Siuly; Li, Yan; Paul Wen, Peng

2014-03-01

Motor imagery (MI) tasks classification provides an important basis for designing brain-computer interface (BCI) systems. If the MI tasks are reliably distinguished through identifying typical patterns in electroencephalography (EEG) data, a motor disabled people could communicate with a device by composing sequences of these mental states. In our earlier study, we developed a cross-correlation based logistic regression (CC-LR) algorithm for the classification of MI tasks for BCI applications, but its performance was not satisfactory. This study develops a modified version of the CC-LR algorithm exploring a suitable feature set that can improve the performance. The modified CC-LR algorithm uses the C3 electrode channel (in the international 10-20 system) as a reference channel for the cross-correlation (CC) technique and applies three diverse feature sets separately, as the input to the logistic regression (LR) classifier. The present algorithm investigates which feature set is the best to characterize the distribution of MI tasks based EEG data. This study also provides an insight into how to select a reference channel for the CC technique with EEG signals considering the anatomical structure of the human brain. The proposed algorithm is compared with eight of the most recently reported well-known methods including the BCI III Winner algorithm. The findings of this study indicate that the modified CC-LR algorithm has potential to improve the identification performance of MI tasks in BCI systems. The results demonstrate that the proposed technique provides a classification improvement over the existing methods tested. Copyright © 2014 Elsevier Ireland Ltd. All rights reserved.

Comparison of Genetic Algorithm, Particle Swarm Optimization and Biogeography-based Optimization for Feature Selection to Classify Clusters of Microcalcifications

NASA Astrophysics Data System (ADS)

Khehra, Baljit Singh; Pharwaha, Amar Partap Singh

2017-04-01

Ductal carcinoma in situ (DCIS) is one type of breast cancer. Clusters of microcalcifications (MCCs) are symptoms of DCIS that are recognized by mammography. Selection of robust features vector is the process of selecting an optimal subset of features from a large number of available features in a given problem domain after the feature extraction and before any classification scheme. Feature selection reduces the feature space that improves the performance of classifier and decreases the computational burden imposed by using many features on classifier. Selection of an optimal subset of features from a large number of available features in a given problem domain is a difficult search problem. For n features, the total numbers of possible subsets of features are 2n. Thus, selection of an optimal subset of features problem belongs to the category of NP-hard problems. In this paper, an attempt is made to find the optimal subset of MCCs features from all possible subsets of features using genetic algorithm (GA), particle swarm optimization (PSO) and biogeography-based optimization (BBO). For simulation, a total of 380 benign and malignant MCCs samples have been selected from mammogram images of DDSM database. A total of 50 features extracted from benign and malignant MCCs samples are used in this study. In these algorithms, fitness function is correct classification rate of classifier. Support vector machine is used as a classifier. From experimental results, it is also observed that the performance of PSO-based and BBO-based algorithms to select an optimal subset of features for classifying MCCs as benign or malignant is better as compared to GA-based algorithm.

Feature Selection Using Information Gain for Improved Structural-Based Alert Correlation

PubMed Central

Siraj, Maheyzah Md; Zainal, Anazida; Elshoush, Huwaida Tagelsir; Elhaj, Fatin

2016-01-01

Grouping and clustering alerts for intrusion detection based on the similarity of features is referred to as structurally base alert correlation and can discover a list of attack steps. Previous researchers selected different features and data sources manually based on their knowledge and experience, which lead to the less accurate identification of attack steps and inconsistent performance of clustering accuracy. Furthermore, the existing alert correlation systems deal with a huge amount of data that contains null values, incomplete information, and irrelevant features causing the analysis of the alerts to be tedious, time-consuming and error-prone. Therefore, this paper focuses on selecting accurate and significant features of alerts that are appropriate to represent the attack steps, thus, enhancing the structural-based alert correlation model. A two-tier feature selection method is proposed to obtain the significant features. The first tier aims at ranking the subset of features based on high information gain entropy in decreasing order. The‏ second tier extends additional features with a better discriminative ability than the initially ranked features. Performance analysis results show the significance of the selected features in terms of the clustering accuracy using 2000 DARPA intrusion detection scenario-specific dataset. PMID:27893821

K-nearest neighbors based methods for identification of different gear crack levels under different motor speeds and loads: Revisited

NASA Astrophysics Data System (ADS)

Wang, Dong

2016-03-01

Gears are the most commonly used components in mechanical transmission systems. Their failures may cause transmission system breakdown and result in economic loss. Identification of different gear crack levels is important to prevent any unexpected gear failure because gear cracks lead to gear tooth breakage. Signal processing based methods mainly require expertize to explain gear fault signatures which is usually not easy to be achieved by ordinary users. In order to automatically identify different gear crack levels, intelligent gear crack identification methods should be developed. The previous case studies experimentally proved that K-nearest neighbors based methods exhibit high prediction accuracies for identification of 3 different gear crack levels under different motor speeds and loads. In this short communication, to further enhance prediction accuracies of existing K-nearest neighbors based methods and extend identification of 3 different gear crack levels to identification of 5 different gear crack levels, redundant statistical features are constructed by using Daubechies 44 (db44) binary wavelet packet transform at different wavelet decomposition levels, prior to the use of a K-nearest neighbors method. The dimensionality of redundant statistical features is 620, which provides richer gear fault signatures. Since many of these statistical features are redundant and highly correlated with each other, dimensionality reduction of redundant statistical features is conducted to obtain new significant statistical features. At last, the K-nearest neighbors method is used to identify 5 different gear crack levels under different motor speeds and loads. A case study including 3 experiments is investigated to demonstrate that the developed method provides higher prediction accuracies than the existing K-nearest neighbors based methods for recognizing different gear crack levels under different motor speeds and loads. Based on the new significant statistical features, some other popular statistical models including linear discriminant analysis, quadratic discriminant analysis, classification and regression tree and naive Bayes classifier, are compared with the developed method. The results show that the developed method has the highest prediction accuracies among these statistical models. Additionally, selection of the number of new significant features and parameter selection of K-nearest neighbors are thoroughly investigated.

A genetic programming approach to oral cancer prognosis.

PubMed

Tan, Mei Sze; Tan, Jing Wei; Chang, Siow-Wee; Yap, Hwa Jen; Abdul Kareem, Sameem; Zain, Rosnah Binti

2016-01-01

The potential of genetic programming (GP) on various fields has been attained in recent years. In bio-medical field, many researches in GP are focused on the recognition of cancerous cells and also on gene expression profiling data. In this research, the aim is to study the performance of GP on the survival prediction of a small sample size of oral cancer prognosis dataset, which is the first study in the field of oral cancer prognosis. GP is applied on an oral cancer dataset that contains 31 cases collected from the Malaysia Oral Cancer Database and Tissue Bank System (MOCDTBS). The feature subsets that is automatically selected through GP were noted and the influences of this subset on the results of GP were recorded. In addition, a comparison between the GP performance and that of the Support Vector Machine (SVM) and logistic regression (LR) are also done in order to verify the predictive capabilities of the GP. The result shows that GP performed the best (average accuracy of 83.87% and average AUROC of 0.8341) when the features selected are smoking, drinking, chewing, histological differentiation of SCC, and oncogene p63. In addition, based on the comparison results, we found that the GP outperformed the SVM and LR in oral cancer prognosis. Some of the features in the dataset are found to be statistically co-related. This is because the accuracy of the GP prediction drops when one of the feature in the best feature subset is excluded. Thus, GP provides an automatic feature selection function, which chooses features that are highly correlated to the prognosis of oral cancer. This makes GP an ideal prediction model for cancer clinical and genomic data that can be used to aid physicians in their decision making stage of diagnosis or prognosis.

Mutual information criterion for feature selection with application to classification of breast microcalcifications

NASA Astrophysics Data System (ADS)

Diamant, Idit; Shalhon, Moran; Goldberger, Jacob; Greenspan, Hayit

2016-03-01

Classification of clustered breast microcalcifications into benign and malignant categories is an extremely challenging task for computerized algorithms and expert radiologists alike. In this paper we present a novel method for feature selection based on mutual information (MI) criterion for automatic classification of microcalcifications. We explored the MI based feature selection for various texture features. The proposed method was evaluated on a standardized digital database for screening mammography (DDSM). Experimental results demonstrate the effectiveness and the advantage of using the MI-based feature selection to obtain the most relevant features for the task and thus to provide for improved performance as compared to using all features.

Regression Levels of Selected Affective Factors on Science Achievement: A Structural Equation Model with TIMSS 2011 Data

ERIC Educational Resources Information Center

Akilli, Mustafa

2015-01-01

The aim of this study is to demonstrate the science success regression levels of chosen emotional features of 8th grade students using Structural Equation Model. The study was conducted by the analysis of students' questionnaires and science success in TIMSS 2011 data using SEM. Initially, the factors that are thought to have an effect on science…

McTwo: a two-step feature selection algorithm based on maximal information coefficient.

PubMed

Ge, Ruiquan; Zhou, Manli; Luo, Youxi; Meng, Qinghan; Mai, Guoqin; Ma, Dongli; Wang, Guoqing; Zhou, Fengfeng

2016-03-23

High-throughput bio-OMIC technologies are producing high-dimension data from bio-samples at an ever increasing rate, whereas the training sample number in a traditional experiment remains small due to various difficulties. This "large p, small n" paradigm in the area of biomedical "big data" may be at least partly solved by feature selection algorithms, which select only features significantly associated with phenotypes. Feature selection is an NP-hard problem. Due to the exponentially increased time requirement for finding the globally optimal solution, all the existing feature selection algorithms employ heuristic rules to find locally optimal solutions, and their solutions achieve different performances on different datasets. This work describes a feature selection algorithm based on a recently published correlation measurement, Maximal Information Coefficient (MIC). The proposed algorithm, McTwo, aims to select features associated with phenotypes, independently of each other, and achieving high classification performance of the nearest neighbor algorithm. Based on the comparative study of 17 datasets, McTwo performs about as well as or better than existing algorithms, with significantly reduced numbers of selected features. The features selected by McTwo also appear to have particular biomedical relevance to the phenotypes from the literature. McTwo selects a feature subset with very good classification performance, as well as a small feature number. So McTwo may represent a complementary feature selection algorithm for the high-dimensional biomedical datasets.

Automatic seed selection for segmentation of liver cirrhosis in laparoscopic sequences

NASA Astrophysics Data System (ADS)

Sinha, Rahul; Marcinczak, Jan Marek; Grigat, Rolf-Rainer

2014-03-01

For computer aided diagnosis based on laparoscopic sequences, image segmentation is one of the basic steps which define the success of all further processing. However, many image segmentation algorithms require prior knowledge which is given by interaction with the clinician. We propose an automatic seed selection algorithm for segmentation of liver cirrhosis in laparoscopic sequences which assigns each pixel a probability of being cirrhotic liver tissue or background tissue. Our approach is based on a trained classifier using SIFT and RGB features with PCA. Due to the unique illumination conditions in laparoscopic sequences of the liver, a very low dimensional feature space can be used for classification via logistic regression. The methodology is evaluated on 718 cirrhotic liver and background patches that are taken from laparoscopic sequences of 7 patients. Using a linear classifier we achieve a precision of 91% in a leave-one-patient-out cross-validation. Furthermore, we demonstrate that with logistic probability estimates, seeds with high certainty of being cirrhotic liver tissue can be obtained. For example, our precision of liver seeds increases to 98.5% if only seeds with more than 95% probability of being liver are used. Finally, these automatically selected seeds can be used as priors in Graph Cuts which is demonstrated in this paper.

Max-AUC Feature Selection in Computer-Aided Detection of Polyps in CT Colonography

PubMed Central

Xu, Jian-Wu; Suzuki, Kenji

2014-01-01

We propose a feature selection method based on a sequential forward floating selection (SFFS) procedure to improve the performance of a classifier in computerized detection of polyps in CT colonography (CTC). The feature selection method is coupled with a nonlinear support vector machine (SVM) classifier. Unlike the conventional linear method based on Wilks' lambda, the proposed method selected the most relevant features that would maximize the area under the receiver operating characteristic curve (AUC), which directly maximizes classification performance, evaluated based on AUC value, in the computer-aided detection (CADe) scheme. We presented two variants of the proposed method with different stopping criteria used in the SFFS procedure. The first variant searched all feature combinations allowed in the SFFS procedure and selected the subsets that maximize the AUC values. The second variant performed a statistical test at each step during the SFFS procedure, and it was terminated if the increase in the AUC value was not statistically significant. The advantage of the second variant is its lower computational cost. To test the performance of the proposed method, we compared it against the popular stepwise feature selection method based on Wilks' lambda for a colonic-polyp database (25 polyps and 2624 nonpolyps). We extracted 75 morphologic, gray-level-based, and texture features from the segmented lesion candidate regions. The two variants of the proposed feature selection method chose 29 and 7 features, respectively. Two SVM classifiers trained with these selected features yielded a 96% by-polyp sensitivity at false-positive (FP) rates of 4.1 and 6.5 per patient, respectively. Experiments showed a significant improvement in the performance of the classifier with the proposed feature selection method over that with the popular stepwise feature selection based on Wilks' lambda that yielded 18.0 FPs per patient at the same sensitivity level. PMID:24608058

Max-AUC feature selection in computer-aided detection of polyps in CT colonography.

PubMed

Xu, Jian-Wu; Suzuki, Kenji

2014-03-01

We propose a feature selection method based on a sequential forward floating selection (SFFS) procedure to improve the performance of a classifier in computerized detection of polyps in CT colonography (CTC). The feature selection method is coupled with a nonlinear support vector machine (SVM) classifier. Unlike the conventional linear method based on Wilks' lambda, the proposed method selected the most relevant features that would maximize the area under the receiver operating characteristic curve (AUC), which directly maximizes classification performance, evaluated based on AUC value, in the computer-aided detection (CADe) scheme. We presented two variants of the proposed method with different stopping criteria used in the SFFS procedure. The first variant searched all feature combinations allowed in the SFFS procedure and selected the subsets that maximize the AUC values. The second variant performed a statistical test at each step during the SFFS procedure, and it was terminated if the increase in the AUC value was not statistically significant. The advantage of the second variant is its lower computational cost. To test the performance of the proposed method, we compared it against the popular stepwise feature selection method based on Wilks' lambda for a colonic-polyp database (25 polyps and 2624 nonpolyps). We extracted 75 morphologic, gray-level-based, and texture features from the segmented lesion candidate regions. The two variants of the proposed feature selection method chose 29 and 7 features, respectively. Two SVM classifiers trained with these selected features yielded a 96% by-polyp sensitivity at false-positive (FP) rates of 4.1 and 6.5 per patient, respectively. Experiments showed a significant improvement in the performance of the classifier with the proposed feature selection method over that with the popular stepwise feature selection based on Wilks' lambda that yielded 18.0 FPs per patient at the same sensitivity level.

BCI Competition IV – Data Set I: Learning Discriminative Patterns for Self-Paced EEG-Based Motor Imagery Detection

PubMed Central

Zhang, Haihong; Guan, Cuntai; Ang, Kai Keng; Wang, Chuanchu

2012-01-01

Detecting motor imagery activities versus non-control in brain signals is the basis of self-paced brain-computer interfaces (BCIs), but also poses a considerable challenge to signal processing due to the complex and non-stationary characteristics of motor imagery as well as non-control. This paper presents a self-paced BCI based on a robust learning mechanism that extracts and selects spatio-spectral features for differentiating multiple EEG classes. It also employs a non-linear regression and post-processing technique for predicting the time-series of class labels from the spatio-spectral features. The method was validated in the BCI Competition IV on Dataset I where it produced the lowest prediction error of class labels continuously. This report also presents and discusses analysis of the method using the competition data set. PMID:22347153

Predicting perceptual quality of images in realistic scenario using deep filter banks

NASA Astrophysics Data System (ADS)

Zhang, Weixia; Yan, Jia; Hu, Shiyong; Ma, Yang; Deng, Dexiang

2018-03-01

Classical image perceptual quality assessment models usually resort to natural scene statistic methods, which are based on an assumption that certain reliable statistical regularities hold on undistorted images and will be corrupted by introduced distortions. However, these models usually fail to accurately predict degradation severity of images in realistic scenarios since complex, multiple, and interactive authentic distortions usually appear on them. We propose a quality prediction model based on convolutional neural network. Quality-aware features extracted from filter banks of multiple convolutional layers are aggregated into the image representation. Furthermore, an easy-to-implement and effective feature selection strategy is used to further refine the image representation and finally a linear support vector regression model is trained to map image representation into images' subjective perceptual quality scores. The experimental results on benchmark databases present the effectiveness and generalizability of the proposed model.

Radiomics analysis of DWI data to identify the rectal cancer patients qualified for local excision after neoadjuvant chemoradiotherapy

NASA Astrophysics Data System (ADS)

Tang, Zhenchao; Liu, Zhenyu; Zhang, Xiaoyan; Shi, Yanjie; Wang, Shou; Fang, Mengjie; Sun, Yingshi; Dong, Enqing; Tian, Jie

2018-02-01

The Locally advanced rectal cancer (LARC) patients were routinely treated with neoadjuvant chemoradiotherapy (CRT) firstly and received total excision afterwards. While, the LARC patients might relieve to T1N0M0/T0N0M0 stage after the CRT, which would enable the patients be qualified for local excision. However, accurate pathological TNM stage could only be obtained by the pathological examination after surgery. We aimed to conduct a Radiomics analysis of Diffusion weighted Imaging (DWI) data to identify the patients in T1N0M0/T0N0M0 stages before surgery, in hope of providing clinical surgery decision support. 223 routinely treated LARC patients in Beijing Cancer Hospital were enrolled in current study. DWI data and clinical characteristics were collected after CRT. According to the pathological TNM stage, the patients of T1N0M0 and T0N0M0 stages were labelled as 1 and the other patients were labelled as 0. The first 123 patients in chronological order were used as training set, and the rest patients as validation set. 563 image features extracted from the DWI data and clinical characteristics were used as features. Two-sample T test was conducted to pre-select the top 50% discriminating features. Least absolute shrinkage and selection operator (Lasso)-Logistic regression model was conducted to further select features and construct the classification model. Based on the 14 selected image features, the area under the Receiver Operating Characteristic (ROC) curve (AUC) of 0.8781, classification Accuracy (ACC) of 0.8432 were achieved in the training set. In the validation set, AUC of 0.8707, ACC (ACC) of 0.84 were observed.

Prediction of Baseflow Index of Catchments using Machine Learning Algorithms

NASA Astrophysics Data System (ADS)

Yadav, B.; Hatfield, K.

2017-12-01

We present the results of eight machine learning techniques for predicting the baseflow index (BFI) of ungauged basins using a surrogate of catchment scale climate and physiographic data. The tested algorithms include ordinary least squares, ridge regression, least absolute shrinkage and selection operator (lasso), elasticnet, support vector machine, gradient boosted regression trees, random forests, and extremely randomized trees. Our work seeks to identify the dominant controls of BFI that can be readily obtained from ancillary geospatial databases and remote sensing measurements, such that the developed techniques can be extended to ungauged catchments. More than 800 gauged catchments spanning the continental United States were selected to develop the general methodology. The BFI calculation was based on the baseflow separated from daily streamflow hydrograph using HYSEP filter. The surrogate catchment attributes were compiled from multiple sources including digital elevation model, soil, landuse, climate data, other publicly available ancillary and geospatial data. 80% catchments were used to train the ML algorithms, and the remaining 20% of the catchments were used as an independent test set to measure the generalization performance of fitted models. A k-fold cross-validation using exhaustive grid search was used to fit the hyperparameters of each model. Initial model development was based on 19 independent variables, but after variable selection and feature ranking, we generated revised sparse models of BFI prediction that are based on only six catchment attributes. These key predictive variables selected after the careful evaluation of bias-variance tradeoff include average catchment elevation, slope, fraction of sand, permeability, temperature, and precipitation. The most promising algorithms exceeding an accuracy score (r-square) of 0.7 on test data include support vector machine, gradient boosted regression trees, random forests, and extremely randomized trees. Considering both the accuracy and the computational complexity of these algorithms, we identify the extremely randomized trees as the best performing algorithm for BFI prediction in ungauged basins.

Development of a stacked ensemble model for forecasting and analyzing daily average PM2.5 concentrations in Beijing, China.

PubMed

Zhai, Binxu; Chen, Jianguo

2018-04-18

A stacked ensemble model is developed for forecasting and analyzing the daily average concentrations of fine particulate matter (PM 2.5 ) in Beijing, China. Special feature extraction procedures, including those of simplification, polynomial, transformation and combination, are conducted before modeling to identify potentially significant features based on an exploratory data analysis. Stability feature selection and tree-based feature selection methods are applied to select important variables and evaluate the degrees of feature importance. Single models including LASSO, Adaboost, XGBoost and multi-layer perceptron optimized by the genetic algorithm (GA-MLP) are established in the level 0 space and are then integrated by support vector regression (SVR) in the level 1 space via stacked generalization. A feature importance analysis reveals that nitrogen dioxide (NO 2 ) and carbon monoxide (CO) concentrations measured from the city of Zhangjiakou are taken as the most important elements of pollution factors for forecasting PM 2.5 concentrations. Local extreme wind speeds and maximal wind speeds are considered to extend the most effects of meteorological factors to the cross-regional transportation of contaminants. Pollutants found in the cities of Zhangjiakou and Chengde have a stronger impact on air quality in Beijing than other surrounding factors. Our model evaluation shows that the ensemble model generally performs better than a single nonlinear forecasting model when applied to new data with a coefficient of determination (R 2 ) of 0.90 and a root mean squared error (RMSE) of 23.69μg/m 3 . For single pollutant grade recognition, the proposed model performs better when applied to days characterized by good air quality than when applied to days registering high levels of pollution. The overall classification accuracy level is 73.93%, with most misclassifications made among adjacent categories. The results demonstrate the interpretability and generalizability of the stacked ensemble model. Copyright © 2018 The Authors. Published by Elsevier B.V. All rights reserved.

Rough sets and Laplacian score based cost-sensitive feature selection

PubMed Central

Yu, Shenglong

2018-01-01

Cost-sensitive feature selection learning is an important preprocessing step in machine learning and data mining. Recently, most existing cost-sensitive feature selection algorithms are heuristic algorithms, which evaluate the importance of each feature individually and select features one by one. Obviously, these algorithms do not consider the relationship among features. In this paper, we propose a new algorithm for minimal cost feature selection called the rough sets and Laplacian score based cost-sensitive feature selection. The importance of each feature is evaluated by both rough sets and Laplacian score. Compared with heuristic algorithms, the proposed algorithm takes into consideration the relationship among features with locality preservation of Laplacian score. We select a feature subset with maximal feature importance and minimal cost when cost is undertaken in parallel, where the cost is given by three different distributions to simulate different applications. Different from existing cost-sensitive feature selection algorithms, our algorithm simultaneously selects out a predetermined number of “good” features. Extensive experimental results show that the approach is efficient and able to effectively obtain the minimum cost subset. In addition, the results of our method are more promising than the results of other cost-sensitive feature selection algorithms. PMID:29912884

Rough sets and Laplacian score based cost-sensitive feature selection.

PubMed

Yu, Shenglong; Zhao, Hong

2018-01-01

Cost-sensitive feature selection learning is an important preprocessing step in machine learning and data mining. Recently, most existing cost-sensitive feature selection algorithms are heuristic algorithms, which evaluate the importance of each feature individually and select features one by one. Obviously, these algorithms do not consider the relationship among features. In this paper, we propose a new algorithm for minimal cost feature selection called the rough sets and Laplacian score based cost-sensitive feature selection. The importance of each feature is evaluated by both rough sets and Laplacian score. Compared with heuristic algorithms, the proposed algorithm takes into consideration the relationship among features with locality preservation of Laplacian score. We select a feature subset with maximal feature importance and minimal cost when cost is undertaken in parallel, where the cost is given by three different distributions to simulate different applications. Different from existing cost-sensitive feature selection algorithms, our algorithm simultaneously selects out a predetermined number of "good" features. Extensive experimental results show that the approach is efficient and able to effectively obtain the minimum cost subset. In addition, the results of our method are more promising than the results of other cost-sensitive feature selection algorithms.

What do you think of my picture? Investigating factors of influence in profile images context perception

NASA Astrophysics Data System (ADS)

Mazza, F.; Da Silva, M. P.; Le Callet, P.; Heynderickx, I. E. J.

2015-03-01

Multimedia quality assessment has been an important research topic during the last decades. The original focus on artifact visibility has been extended during the years to aspects as image aesthetics, interestingness and memorability. More recently, Fedorovskaya proposed the concept of 'image psychology': this concept focuses on additional quality dimensions related to human content processing. While these additional dimensions are very valuable in understanding preferences, it is very hard to define, isolate and measure their effect on quality. In this paper we continue our research on face pictures investigating which image factors influence context perception. We collected perceived fit of a set of images to various content categories. These categories were selected based on current typologies in social networks. Logistic regression was adopted to model category fit based on images features. In this model we used both low level and high level features, the latter focusing on complex features related to image content. In order to extract these high level features, we relied on crowdsourcing, since computer vision algorithms are not yet sufficiently accurate for the features we needed. Our results underline the importance of some high level content features, e.g. the dress of the portrayed person and scene setting, in categorizing image.

Quantification of dead vegetation fraction in mixed pastures using AisaFENIX imaging spectroscopy data

NASA Astrophysics Data System (ADS)

Pullanagari, R. R.; Kereszturi, G.; Yule, I. J.

2017-06-01

New Zealand farming relies heavily on grazed pasture for feeding livestock; therefore it is important to provide high quality palatable grass in order to maintain profitable and sustainable grassland management. The presence of non-photosynthetic vegetation (NPV) such as dead vegetation in pastures severely limits the quality and productivity of pastures. Quantifying the fraction of dead vegetation in mixed pastures is a great challenge even with remote sensing approaches. In this study, a high spatial resolution with pixel resolution of 1 m and spectral resolution of 3.5-5.6 nm imaging spectroscopy data from AisaFENIX (380-2500 nm) was used to assess the fraction of dead vegetation component in mixed pastures on a hill country farm in New Zealand. We used different methods to retrieve dead vegetation fraction from the spectra; narrow band vegetation indices, full spectrum based partial least squares (PLS) regression and feature selection based PLS regression. Among all approaches, feature selection based PLS model exhibited better performance in terms of prediction accuracy (R2CV = 0.73, RMSECV = 6.05, RPDCV = 2.25). The results were consistent with validation data, and also performed well on the external test data (R2 = 0.62, RMSE = 8.06, RPD = 2.06). In addition, statistical tests were conducted to ascertain the effect of topographical variables such as slope and aspect on the accumulation of the dead vegetation fraction. Steep slopes (>25°) had a significantly (p < 0.05) higher amount of dead vegetation. In contrast, aspect showed non-significant impact on dead vegetation accumulation. The results from the study indicate that AisaFENIX imaging spectroscopy data could be a useful tool for mapping the dead vegetation fraction accurately.

Speech Emotion Feature Selection Method Based on Contribution Analysis Algorithm of Neural Network

DOE Office of Scientific and Technical Information (OSTI.GOV)

Wang Xiaojia; Mao Qirong; Zhan Yongzhao

There are many emotion features. If all these features are employed to recognize emotions, redundant features may be existed. Furthermore, recognition result is unsatisfying and the cost of feature extraction is high. In this paper, a method to select speech emotion features based on contribution analysis algorithm of NN is presented. The emotion features are selected by using contribution analysis algorithm of NN from the 95 extracted features. Cluster analysis is applied to analyze the effectiveness for the features selected, and the time of feature extraction is evaluated. Finally, 24 emotion features selected are used to recognize six speech emotions.more » The experiments show that this method can improve the recognition rate and the time of feature extraction.« less

Use of a Machine-learning Method for Predicting Highly Cited Articles Within General Radiology Journals.

PubMed

Rosenkrantz, Andrew B; Doshi, Ankur M; Ginocchio, Luke A; Aphinyanaphongs, Yindalon

2016-12-01

This study aimed to assess the performance of a text classification machine-learning model in predicting highly cited articles within the recent radiological literature and to identify the model's most influential article features. We downloaded from PubMed the title, abstract, and medical subject heading terms for 10,065 articles published in 25 general radiology journals in 2012 and 2013. Three machine-learning models were applied to predict the top 10% of included articles in terms of the number of citations to the article in 2014 (reflecting the 2-year time window in conventional impact factor calculations). The model having the highest area under the curve was selected to derive a list of article features (words) predicting high citation volume, which was iteratively reduced to identify the smallest possible core feature list maintaining predictive power. Overall themes were qualitatively assigned to the core features. The regularized logistic regression (Bayesian binary regression) model had highest performance, achieving an area under the curve of 0.814 in predicting articles in the top 10% of citation volume. We reduced the initial 14,083 features to 210 features that maintain predictivity. These features corresponded with topics relating to various imaging techniques (eg, diffusion-weighted magnetic resonance imaging, hyperpolarized magnetic resonance imaging, dual-energy computed tomography, computed tomography reconstruction algorithms, tomosynthesis, elastography, and computer-aided diagnosis), particular pathologies (prostate cancer; thyroid nodules; hepatic adenoma, hepatocellular carcinoma, non-alcoholic fatty liver disease), and other topics (radiation dose, electroporation, education, general oncology, gadolinium, statistics). Machine learning can be successfully applied to create specific feature-based models for predicting articles likely to achieve high influence within the radiological literature. Copyright © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.

SU-F-R-22: Malignancy Classification for Small Pulmonary Nodules with Radiomics and Logistic Regression

DOE Office of Scientific and Technical Information (OSTI.GOV)

Huang, W; Tu, S

Purpose: We conducted a retrospective study of Radiomics research for classifying malignancy of small pulmonary nodules. A machine learning algorithm of logistic regression and open research platform of Radiomics, IBEX (Imaging Biomarker Explorer), were used to evaluate the classification accuracy. Methods: The training set included 100 CT image series from cancer patients with small pulmonary nodules where the average diameter is 1.10 cm. These patients registered at Chang Gung Memorial Hospital and received a CT-guided operation of lung cancer lobectomy. The specimens were classified by experienced pathologists with a B (benign) or M (malignant). CT images with slice thickness ofmore » 0.625 mm were acquired from a GE BrightSpeed 16 scanner. The study was formally approved by our institutional internal review board. Nodules were delineated and 374 feature parameters were extracted from IBEX. We first used the t-test and p-value criteria to study which feature can differentiate between group B and M. Then we implemented a logistic regression algorithm to perform nodule malignancy classification. 10-fold cross-validation and the receiver operating characteristic curve (ROC) were used to evaluate the classification accuracy. Finally hierarchical clustering analysis, Spearman rank correlation coefficient, and clustering heat map were used to further study correlation characteristics among different features. Results: 238 features were found differentiable between group B and M based on whether their statistical p-values were less than 0.05. A forward search algorithm was used to select an optimal combination of features for the best classification and 9 features were identified. Our study found the best accuracy of classifying malignancy was 0.79±0.01 with the 10-fold cross-validation. The area under the ROC curve was 0.81±0.02. Conclusion: Benign nodules may be treated as a malignant tumor in low-dose CT and patients may undergo unnecessary surgeries or treatments. Our study may help radiologists to differentiate nodule malignancy for low-dose CT.« less

Improving Classification of Protein Interaction Articles Using Context Similarity-Based Feature Selection.

PubMed

Chen, Yifei; Sun, Yuxing; Han, Bing-Qing

2015-01-01

Protein interaction article classification is a text classification task in the biological domain to determine which articles describe protein-protein interactions. Since the feature space in text classification is high-dimensional, feature selection is widely used for reducing the dimensionality of features to speed up computation without sacrificing classification performance. Many existing feature selection methods are based on the statistical measure of document frequency and term frequency. One potential drawback of these methods is that they treat features separately. Hence, first we design a similarity measure between the context information to take word cooccurrences and phrase chunks around the features into account. Then we introduce the similarity of context information to the importance measure of the features to substitute the document and term frequency. Hence we propose new context similarity-based feature selection methods. Their performance is evaluated on two protein interaction article collections and compared against the frequency-based methods. The experimental results reveal that the context similarity-based methods perform better in terms of the F1 measure and the dimension reduction rate. Benefiting from the context information surrounding the features, the proposed methods can select distinctive features effectively for protein interaction article classification.

Learning templates for artistic portrait lighting analysis.

PubMed

Chen, Xiaowu; Jin, Xin; Wu, Hongyu; Zhao, Qinping

2015-02-01

Lighting is a key factor in creating impressive artistic portraits. In this paper, we propose to analyze portrait lighting by learning templates of lighting styles. Inspired by the experience of artists, we first define several novel features that describe the local contrasts in various face regions. The most informative features are then selected with a stepwise feature pursuit algorithm to derive the templates of various lighting styles. After that, the matching scores that measure the similarity between a testing portrait and those templates are calculated for lighting style classification. Furthermore, we train a regression model by the subjective scores and the feature responses of a template to predict the score of a portrait lighting quality. Based on the templates, a novel face illumination descriptor is defined to measure the difference between two portrait lightings. Experimental results show that the learned templates can well describe the lighting styles, whereas the proposed approach can assess the lighting quality of artistic portraits as human being does.

Discrimination of serum Raman spectroscopy between normal and colorectal cancer

NASA Astrophysics Data System (ADS)

Li, Xiaozhou; Yang, Tianyue; Yu, Ting; Li, Siqi

2011-07-01

Raman spectroscopy of tissues has been widely studied for the diagnosis of various cancers, but biofluids were seldom used as the analyte because of the low concentration. Herein, serum of 30 normal people, 46 colon cancer, and 44 rectum cancer patients were measured Raman spectra and analyzed. The information of Raman peaks (intensity and width) and that of the fluorescence background (baseline function coefficients) were selected as parameters for statistical analysis. Principal component regression (PCR) and partial least square regression (PLSR) were used on the selected parameters separately to see the performance of the parameters. PCR performed better than PLSR in our spectral data. Then linear discriminant analysis (LDA) was used on the principal components (PCs) of the two regression method on the selected parameters, and a diagnostic accuracy of 88% and 83% were obtained. The conclusion is that the selected features can maintain the information of original spectra well and Raman spectroscopy of serum has the potential for the diagnosis of colorectal cancer.

Feature Selection in Classification of Eye Movements Using Electrooculography for Activity Recognition

PubMed Central

Mala, S.; Latha, K.

2014-01-01

Activity recognition is needed in different requisition, for example, reconnaissance system, patient monitoring, and human-computer interfaces. Feature selection plays an important role in activity recognition, data mining, and machine learning. In selecting subset of features, an efficient evolutionary algorithm Differential Evolution (DE), a very efficient optimizer, is used for finding informative features from eye movements using electrooculography (EOG). Many researchers use EOG signals in human-computer interactions with various computational intelligence methods to analyze eye movements. The proposed system involves analysis of EOG signals using clearness based features, minimum redundancy maximum relevance features, and Differential Evolution based features. This work concentrates more on the feature selection algorithm based on DE in order to improve the classification for faultless activity recognition. PMID:25574185

Feature selection in classification of eye movements using electrooculography for activity recognition.

PubMed

Mala, S; Latha, K

2014-01-01

Activity recognition is needed in different requisition, for example, reconnaissance system, patient monitoring, and human-computer interfaces. Feature selection plays an important role in activity recognition, data mining, and machine learning. In selecting subset of features, an efficient evolutionary algorithm Differential Evolution (DE), a very efficient optimizer, is used for finding informative features from eye movements using electrooculography (EOG). Many researchers use EOG signals in human-computer interactions with various computational intelligence methods to analyze eye movements. The proposed system involves analysis of EOG signals using clearness based features, minimum redundancy maximum relevance features, and Differential Evolution based features. This work concentrates more on the feature selection algorithm based on DE in order to improve the classification for faultless activity recognition.

Developing a NIR multispectral imaging for prediction and visualization of peanut protein content using variable selection algorithms

NASA Astrophysics Data System (ADS)

Cheng, Jun-Hu; Jin, Huali; Liu, Zhiwei

2018-01-01

The feasibility of developing a multispectral imaging method using important wavelengths from hyperspectral images selected by genetic algorithm (GA), successive projection algorithm (SPA) and regression coefficient (RC) methods for modeling and predicting protein content in peanut kernel was investigated for the first time. Partial least squares regression (PLSR) calibration model was established between the spectral data from the selected optimal wavelengths and the reference measured protein content ranged from 23.46% to 28.43%. The RC-PLSR model established using eight key wavelengths (1153, 1567, 1972, 2143, 2288, 2339, 2389 and 2446 nm) showed the best predictive results with the coefficient of determination of prediction (R2P) of 0.901, and root mean square error of prediction (RMSEP) of 0.108 and residual predictive deviation (RPD) of 2.32. Based on the obtained best model and image processing algorithms, the distribution maps of protein content were generated. The overall results of this study indicated that developing a rapid and online multispectral imaging system using the feature wavelengths and PLSR analysis is potential and feasible for determination of the protein content in peanut kernels.

New variable selection methods for zero-inflated count data with applications to the substance abuse field

PubMed Central

Buu, Anne; Johnson, Norman J.; Li, Runze; Tan, Xianming

2011-01-01

Zero-inflated count data are very common in health surveys. This study develops new variable selection methods for the zero-inflated Poisson regression model. Our simulations demonstrate the negative consequences which arise from the ignorance of zero-inflation. Among the competing methods, the one-step SCAD method is recommended because it has the highest specificity, sensitivity, exact fit, and lowest estimation error. The design of the simulations is based on the special features of two large national databases commonly used in the alcoholism and substance abuse field so that our findings can be easily generalized to the real settings. Applications of the methodology are demonstrated by empirical analyses on the data from a well-known alcohol study. PMID:21563207

Recognizing Banknote Fitness with a Visible Light One Dimensional Line Image Sensor

PubMed Central

Pham, Tuyen Danh; Park, Young Ho; Kwon, Seung Yong; Nguyen, Dat Tien; Vokhidov, Husan; Park, Kang Ryoung; Jeong, Dae Sik; Yoon, Sungsoo

2015-01-01

In general, dirty banknotes that have creases or soiled surfaces should be replaced by new banknotes, whereas clean banknotes should be recirculated. Therefore, the accurate classification of banknote fitness when sorting paper currency is an important and challenging task. Most previous research has focused on sensors that used visible, infrared, and ultraviolet light. Furthermore, there was little previous research on the fitness classification for Indian paper currency. Therefore, we propose a new method for classifying the fitness of Indian banknotes, with a one-dimensional line image sensor that uses only visible light. The fitness of banknotes is usually determined by various factors such as soiling, creases, and tears, etc. although we just consider banknote soiling in our research. This research is novel in the following four ways: first, there has been little research conducted on fitness classification for the Indian Rupee using visible-light images. Second, the classification is conducted based on the features extracted from the regions of interest (ROIs), which contain little texture. Third, 1-level discrete wavelet transformation (DWT) is used to extract the features for discriminating between fit and unfit banknotes. Fourth, the optimal DWT features that represent the fitness and unfitness of banknotes are selected based on linear regression analysis with ground-truth data measured by densitometer. In addition, the selected features are used as the inputs to a support vector machine (SVM) for the final classification of banknote fitness. Experimental results showed that our method outperforms other methods. PMID:26343654

Sequence Based Prediction of DNA-Binding Proteins Based on Hybrid Feature Selection Using Random Forest and Gaussian Naïve Bayes

PubMed Central

Lou, Wangchao; Wang, Xiaoqing; Chen, Fan; Chen, Yixiao; Jiang, Bo; Zhang, Hua

2014-01-01

Developing an efficient method for determination of the DNA-binding proteins, due to their vital roles in gene regulation, is becoming highly desired since it would be invaluable to advance our understanding of protein functions. In this study, we proposed a new method for the prediction of the DNA-binding proteins, by performing the feature rank using random forest and the wrapper-based feature selection using forward best-first search strategy. The features comprise information from primary sequence, predicted secondary structure, predicted relative solvent accessibility, and position specific scoring matrix. The proposed method, called DBPPred, used Gaussian naïve Bayes as the underlying classifier since it outperformed five other classifiers, including decision tree, logistic regression, k-nearest neighbor, support vector machine with polynomial kernel, and support vector machine with radial basis function. As a result, the proposed DBPPred yields the highest average accuracy of 0.791 and average MCC of 0.583 according to the five-fold cross validation with ten runs on the training benchmark dataset PDB594. Subsequently, blind tests on the independent dataset PDB186 by the proposed model trained on the entire PDB594 dataset and by other five existing methods (including iDNA-Prot, DNA-Prot, DNAbinder, DNABIND and DBD-Threader) were performed, resulting in that the proposed DBPPred yielded the highest accuracy of 0.769, MCC of 0.538, and AUC of 0.790. The independent tests performed by the proposed DBPPred on completely a large non-DNA binding protein dataset and two RNA binding protein datasets also showed improved or comparable quality when compared with the relevant prediction methods. Moreover, we observed that majority of the selected features by the proposed method are statistically significantly different between the mean feature values of the DNA-binding and the non DNA-binding proteins. All of the experimental results indicate that the proposed DBPPred can be an alternative perspective predictor for large-scale determination of DNA-binding proteins. PMID:24475169

Vowel Imagery Decoding toward Silent Speech BCI Using Extreme Learning Machine with Electroencephalogram

PubMed Central

Kim, Jongin; Park, Hyeong-jun

2016-01-01

The purpose of this study is to classify EEG data on imagined speech in a single trial. We recorded EEG data while five subjects imagined different vowels, /a/, /e/, /i/, /o/, and /u/. We divided each single trial dataset into thirty segments and extracted features (mean, variance, standard deviation, and skewness) from all segments. To reduce the dimension of the feature vector, we applied a feature selection algorithm based on the sparse regression model. These features were classified using a support vector machine with a radial basis function kernel, an extreme learning machine, and two variants of an extreme learning machine with different kernels. Because each single trial consisted of thirty segments, our algorithm decided the label of the single trial by selecting the most frequent output among the outputs of the thirty segments. As a result, we observed that the extreme learning machine and its variants achieved better classification rates than the support vector machine with a radial basis function kernel and linear discrimination analysis. Thus, our results suggested that EEG responses to imagined speech could be successfully classified in a single trial using an extreme learning machine with a radial basis function and linear kernel. This study with classification of imagined speech might contribute to the development of silent speech BCI systems. PMID:28097128

Differential evolution-based multi-objective optimization for the definition of a health indicator for fault diagnostics and prognostics

NASA Astrophysics Data System (ADS)

Baraldi, P.; Bonfanti, G.; Zio, E.

2018-03-01

The identification of the current degradation state of an industrial component and the prediction of its future evolution is a fundamental step for the development of condition-based and predictive maintenance approaches. The objective of the present work is to propose a general method for extracting a health indicator to measure the amount of component degradation from a set of signals measured during operation. The proposed method is based on the combined use of feature extraction techniques, such as Empirical Mode Decomposition and Auto-Associative Kernel Regression, and a multi-objective Binary Differential Evolution (BDE) algorithm for selecting the subset of features optimal for the definition of the health indicator. The objectives of the optimization are desired characteristics of the health indicator, such as monotonicity, trendability and prognosability. A case study is considered, concerning the prediction of the remaining useful life of turbofan engines. The obtained results confirm that the method is capable of extracting health indicators suitable for accurate prognostics.

Feature selection method based on multi-fractal dimension and harmony search algorithm and its application

NASA Astrophysics Data System (ADS)

Zhang, Chen; Ni, Zhiwei; Ni, Liping; Tang, Na

2016-10-01

Feature selection is an important method of data preprocessing in data mining. In this paper, a novel feature selection method based on multi-fractal dimension and harmony search algorithm is proposed. Multi-fractal dimension is adopted as the evaluation criterion of feature subset, which can determine the number of selected features. An improved harmony search algorithm is used as the search strategy to improve the efficiency of feature selection. The performance of the proposed method is compared with that of other feature selection algorithms on UCI data-sets. Besides, the proposed method is also used to predict the daily average concentration of PM2.5 in China. Experimental results show that the proposed method can obtain competitive results in terms of both prediction accuracy and the number of selected features.

Bayesian quantile regression-based partially linear mixed-effects joint models for longitudinal data with multiple features.

PubMed

Zhang, Hanze; Huang, Yangxin; Wang, Wei; Chen, Henian; Langland-Orban, Barbara

2017-01-01

In longitudinal AIDS studies, it is of interest to investigate the relationship between HIV viral load and CD4 cell counts, as well as the complicated time effect. Most of common models to analyze such complex longitudinal data are based on mean-regression, which fails to provide efficient estimates due to outliers and/or heavy tails. Quantile regression-based partially linear mixed-effects models, a special case of semiparametric models enjoying benefits of both parametric and nonparametric models, have the flexibility to monitor the viral dynamics nonparametrically and detect the varying CD4 effects parametrically at different quantiles of viral load. Meanwhile, it is critical to consider various data features of repeated measurements, including left-censoring due to a limit of detection, covariate measurement error, and asymmetric distribution. In this research, we first establish a Bayesian joint models that accounts for all these data features simultaneously in the framework of quantile regression-based partially linear mixed-effects models. The proposed models are applied to analyze the Multicenter AIDS Cohort Study (MACS) data. Simulation studies are also conducted to assess the performance of the proposed methods under different scenarios.

Toward high-throughput phenotyping: unbiased automated feature extraction and selection from knowledge sources.

PubMed

Yu, Sheng; Liao, Katherine P; Shaw, Stanley Y; Gainer, Vivian S; Churchill, Susanne E; Szolovits, Peter; Murphy, Shawn N; Kohane, Isaac S; Cai, Tianxi

2015-09-01

Analysis of narrative (text) data from electronic health records (EHRs) can improve population-scale phenotyping for clinical and genetic research. Currently, selection of text features for phenotyping algorithms is slow and laborious, requiring extensive and iterative involvement by domain experts. This paper introduces a method to develop phenotyping algorithms in an unbiased manner by automatically extracting and selecting informative features, which can be comparable to expert-curated ones in classification accuracy. Comprehensive medical concepts were collected from publicly available knowledge sources in an automated, unbiased fashion. Natural language processing (NLP) revealed the occurrence patterns of these concepts in EHR narrative notes, which enabled selection of informative features for phenotype classification. When combined with additional codified features, a penalized logistic regression model was trained to classify the target phenotype. The authors applied our method to develop algorithms to identify patients with rheumatoid arthritis and coronary artery disease cases among those with rheumatoid arthritis from a large multi-institutional EHR. The area under the receiver operating characteristic curves (AUC) for classifying RA and CAD using models trained with automated features were 0.951 and 0.929, respectively, compared to the AUCs of 0.938 and 0.929 by models trained with expert-curated features. Models trained with NLP text features selected through an unbiased, automated procedure achieved comparable or slightly higher accuracy than those trained with expert-curated features. The majority of the selected model features were interpretable. The proposed automated feature extraction method, generating highly accurate phenotyping algorithms with improved efficiency, is a significant step toward high-throughput phenotyping. © The Author 2015. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

Integrating Dimension Reduction and Out-of-Sample Extension in Automated Classification of Ex Vivo Human Patellar Cartilage on Phase Contrast X-Ray Computed Tomography

PubMed Central

Nagarajan, Mahesh B.; Coan, Paola; Huber, Markus B.; Diemoz, Paul C.; Wismüller, Axel

2015-01-01

Phase contrast X-ray computed tomography (PCI-CT) has been demonstrated as a novel imaging technique that can visualize human cartilage with high spatial resolution and soft tissue contrast. Different textural approaches have been previously investigated for characterizing chondrocyte organization on PCI-CT to enable classification of healthy and osteoarthritic cartilage. However, the large size of feature sets extracted in such studies motivates an investigation into algorithmic feature reduction for computing efficient feature representations without compromising their discriminatory power. For this purpose, geometrical feature sets derived from the scaling index method (SIM) were extracted from 1392 volumes of interest (VOI) annotated on PCI-CT images of ex vivo human patellar cartilage specimens. The extracted feature sets were subject to linear and non-linear dimension reduction techniques as well as feature selection based on evaluation of mutual information criteria. The reduced feature set was subsequently used in a machine learning task with support vector regression to classify VOIs as healthy or osteoarthritic; classification performance was evaluated using the area under the receiver-operating characteristic (ROC) curve (AUC). Our results show that the classification performance achieved by 9-D SIM-derived geometric feature sets (AUC: 0.96 ± 0.02) can be maintained with 2-D representations computed from both dimension reduction and feature selection (AUC values as high as 0.97 ± 0.02). Thus, such feature reduction techniques can offer a high degree of compaction to large feature sets extracted from PCI-CT images while maintaining their ability to characterize the underlying chondrocyte patterns. PMID:25710875

Reducing the number of reconstructions needed for estimating channelized observer performance

NASA Astrophysics Data System (ADS)

Pineda, Angel R.; Miedema, Hope; Brenner, Melissa; Altaf, Sana

2018-03-01

A challenge for task-based optimization is the time required for each reconstructed image in applications where reconstructions are time consuming. Our goal is to reduce the number of reconstructions needed to estimate the area under the receiver operating characteristic curve (AUC) of the infinitely-trained optimal channelized linear observer. We explore the use of classifiers which either do not invert the channel covariance matrix or do feature selection. We also study the assumption that multiple low contrast signals in the same image of a non-linear reconstruction do not significantly change the estimate of the AUC. We compared the AUC of several classifiers (Hotelling, logistic regression, logistic regression using Firth bias reduction and the least absolute shrinkage and selection operator (LASSO)) with a small number of observations both for normal simulated data and images from a total variation reconstruction in magnetic resonance imaging (MRI). We used 10 Laguerre-Gauss channels and the Mann-Whitney estimator for AUC. For this data, our results show that at small sample sizes feature selection using the LASSO technique can decrease bias of the AUC estimation with increased variance and that for large sample sizes the difference between these classifiers is small. We also compared the use of multiple signals in a single reconstructed image to reduce the number of reconstructions in a total variation reconstruction for accelerated imaging in MRI. We found that AUC estimation using multiple low contrast signals in the same image resulted in similar AUC estimates as doing a single reconstruction per signal leading to a 13x reduction in the number of reconstructions needed.

Finite-sample and asymptotic sign-based tests for parameters of non-linear quantile regression with Markov noise

NASA Astrophysics Data System (ADS)

Sirenko, M. A.; Tarasenko, P. F.; Pushkarev, M. I.

2017-01-01

One of the most noticeable features of sign-based statistical procedures is an opportunity to build an exact test for simple hypothesis testing of parameters in a regression model. In this article, we expanded a sing-based approach to the nonlinear case with dependent noise. The examined model is a multi-quantile regression, which makes it possible to test hypothesis not only of regression parameters, but of noise parameters as well.

Development of an automated ultrasonic testing system

NASA Astrophysics Data System (ADS)

Shuxiang, Jiao; Wong, Brian Stephen

2005-04-01

Non-Destructive Testing is necessary in areas where defects in structures emerge over time due to wear and tear and structural integrity is necessary to maintain its usability. However, manual testing results in many limitations: high training cost, long training procedure, and worse, the inconsistent test results. A prime objective of this project is to develop an automatic Non-Destructive testing system for a shaft of the wheel axle of a railway carriage. Various methods, such as the neural network, pattern recognition methods and knowledge-based system are used for the artificial intelligence problem. In this paper, a statistical pattern recognition approach, Classification Tree is applied. Before feature selection, a thorough study on the ultrasonic signals produced was carried out. Based on the analysis of the ultrasonic signals, three signal processing methods were developed to enhance the ultrasonic signals: Cross-Correlation, Zero-Phase filter and Averaging. The target of this step is to reduce the noise and make the signal character more distinguishable. Four features: 1. The Auto Regressive Model Coefficients. 2. Standard Deviation. 3. Pearson Correlation 4. Dispersion Uniformity Degree are selected. And then a Classification Tree is created and applied to recognize the peak positions and amplitudes. Searching local maximum is carried out before feature computing. This procedure reduces much computation time in the real-time testing. Based on this algorithm, a software package called SOFRA was developed to recognize the peaks, calibrate automatically and test a simulated shaft automatically. The automatic calibration procedure and the automatic shaft testing procedure are developed.

An evaluation of supervised classifiers for indirectly detecting salt-affected areas at irrigation scheme level

NASA Astrophysics Data System (ADS)

Muller, Sybrand Jacobus; van Niekerk, Adriaan

2016-07-01

Soil salinity often leads to reduced crop yield and quality and can render soils barren. Irrigated areas are particularly at risk due to intensive cultivation and secondary salinization caused by waterlogging. Regular monitoring of salt accumulation in irrigation schemes is needed to keep its negative effects under control. The dynamic spatial and temporal characteristics of remote sensing can provide a cost-effective solution for monitoring salt accumulation at irrigation scheme level. This study evaluated a range of pan-fused SPOT-5 derived features (spectral bands, vegetation indices, image textures and image transformations) for classifying salt-affected areas in two distinctly different irrigation schemes in South Africa, namely Vaalharts and Breede River. The relationship between the input features and electro conductivity measurements were investigated using regression modelling (stepwise linear regression, partial least squares regression, curve fit regression modelling) and supervised classification (maximum likelihood, nearest neighbour, decision tree analysis, support vector machine and random forests). Classification and regression trees and random forest were used to select the most important features for differentiating salt-affected and unaffected areas. The results showed that the regression analyses produced weak models (<0.4 R squared). Better results were achieved using the supervised classifiers, but the algorithms tend to over-estimate salt-affected areas. A key finding was that none of the feature sets or classification algorithms stood out as being superior for monitoring salt accumulation at irrigation scheme level. This was attributed to the large variations in the spectral responses of different crops types at different growing stages, coupled with their individual tolerances to saline conditions.

TU-AB-BRA-10: Prognostic Value of Intra-Radiation Treatment FDG-PET and CT Imaging Features in Locally Advanced Head and Neck Cancer

DOE Office of Scientific and Technical Information (OSTI.GOV)

Song, J; Pollom, E; Durkee, B

2015-06-15

Purpose: To predict response to radiation treatment using computational FDG-PET and CT images in locally advanced head and neck cancer (HNC). Methods: 68 patients with State III-IVB HNC treated with chemoradiation were included in this retrospective study. For each patient, we analyzed primary tumor and lymph nodes on PET and CT scans acquired both prior to and during radiation treatment, which led to 8 combinations of image datasets. From each image set, we extracted high-throughput, radiomic features of the following types: statistical, morphological, textural, histogram, and wavelet, resulting in a total of 437 features. We then performed unsupervised redundancy removalmore » and stability test on these features. To avoid over-fitting, we trained a logistic regression model with simultaneous feature selection based on least absolute shrinkage and selection operator (LASSO). To objectively evaluate the prediction ability, we performed 5-fold cross validation (CV) with 50 random repeats of stratified bootstrapping. Feature selection and model training was solely conducted on the training set and independently validated on the holdout test set. Receiver operating characteristic (ROC) curve of the pooled Result and the area under the ROC curve (AUC) was calculated as figure of merit. Results: For predicting local-regional recurrence, our model built on pre-treatment PET of lymph nodes achieved the best performance (AUC=0.762) on 5-fold CV, which compared favorably with node volume and SUVmax (AUC=0.704 and 0.449, p<0.001). Wavelet coefficients turned out to be the most predictive features. Prediction of distant recurrence showed a similar trend, in which pre-treatment PET features of lymph nodes had the highest AUC of 0.705. Conclusion: The radiomics approach identified novel imaging features that are predictive to radiation treatment response. If prospectively validated in larger cohorts, they could aid in risk-adaptive treatment of HNC.« less

Skin lesion computational diagnosis of dermoscopic images: Ensemble models based on input feature manipulation.

PubMed

Oliveira, Roberta B; Pereira, Aledir S; Tavares, João Manuel R S

2017-10-01

The number of deaths worldwide due to melanoma has risen in recent times, in part because melanoma is the most aggressive type of skin cancer. Computational systems have been developed to assist dermatologists in early diagnosis of skin cancer, or even to monitor skin lesions. However, there still remains a challenge to improve classifiers for the diagnosis of such skin lesions. The main objective of this article is to evaluate different ensemble classification models based on input feature manipulation to diagnose skin lesions. Input feature manipulation processes are based on feature subset selections from shape properties, colour variation and texture analysis to generate diversity for the ensemble models. Three subset selection models are presented here: (1) a subset selection model based on specific feature groups, (2) a correlation-based subset selection model, and (3) a subset selection model based on feature selection algorithms. Each ensemble classification model is generated using an optimum-path forest classifier and integrated with a majority voting strategy. The proposed models were applied on a set of 1104 dermoscopic images using a cross-validation procedure. The best results were obtained by the first ensemble classification model that generates a feature subset ensemble based on specific feature groups. The skin lesion diagnosis computational system achieved 94.3% accuracy, 91.8% sensitivity and 96.7% specificity. The input feature manipulation process based on specific feature subsets generated the greatest diversity for the ensemble classification model with very promising results. Copyright © 2017 Elsevier B.V. All rights reserved.

Attentional Selection of Feature Conjunctions Is Accomplished by Parallel and Independent Selection of Single Features.

PubMed

Andersen, Søren K; Müller, Matthias M; Hillyard, Steven A

2015-07-08

Experiments that study feature-based attention have often examined situations in which selection is based on a single feature (e.g., the color red). However, in more complex situations relevant stimuli may not be set apart from other stimuli by a single defining property but by a specific combination of features. Here, we examined sustained attentional selection of stimuli defined by conjunctions of color and orientation. Human observers attended to one out of four concurrently presented superimposed fields of randomly moving horizontal or vertical bars of red or blue color to detect brief intervals of coherent motion. Selective stimulus processing in early visual cortex was assessed by recordings of steady-state visual evoked potentials (SSVEPs) elicited by each of the flickering fields of stimuli. We directly contrasted attentional selection of single features and feature conjunctions and found that SSVEP amplitudes on conditions in which selection was based on a single feature only (color or orientation) exactly predicted the magnitude of attentional enhancement of SSVEPs when attending to a conjunction of both features. Furthermore, enhanced SSVEP amplitudes elicited by attended stimuli were accompanied by equivalent reductions of SSVEP amplitudes elicited by unattended stimuli in all cases. We conclude that attentional selection of a feature-conjunction stimulus is accomplished by the parallel and independent facilitation of its constituent feature dimensions in early visual cortex. The ability to perceive the world is limited by the brain's processing capacity. Attention affords adaptive behavior by selectively prioritizing processing of relevant stimuli based on their features (location, color, orientation, etc.). We found that attentional mechanisms for selection of different features belonging to the same object operate independently and in parallel: concurrent attentional selection of two stimulus features is simply the sum of attending to each of those features separately. This result is key to understanding attentional selection in complex (natural) scenes, where relevant stimuli are likely to be defined by a combination of stimulus features. Copyright © 2015 the authors 0270-6474/15/359912-08$15.00/0.

Applying dam height-storage curve to geomorphic features analysis within virtual geographic environment: a case study of the Hong-Shi-Mao watershed

NASA Astrophysics Data System (ADS)

Wang, Daojun; Gong, Jianhua; Ma, Ainai; Li, Wenhang; Wang, Xijun

2005-10-01

There are generally two kinds of approaches to studying geomorphic features in terms of the quantification level and difference of major considerations. One is the earlier qualitative characterization, and the other is the 2-dimension measurement that includes section pattern and projection pattern. With the development of geo-information technology, especially the 3-D geo-visualization and virtual geographic environments (VGE), 3-dimension measurement and dynamic interactive between users and geo-data/geo-graphics can be developed to understand geomorphic features deeply, and to benefit to the effective applications of such features for geographic projects like dam construction. Storage-elevation curve is very useful for site selection of projects and flood dispatching in water conservancy region, but it is just a tool querying one value from the other one. In fact, storage-elevation curve can represent comprehensively the geomorphic features including vertical section, cross section of the stream and the landform nearby. In this paper, we use quadratic regression equation shaped like y = ax2 + bx + c and the DEM data of Hong-Shi-Mao watershed, Zi Chang County, ShaanXi Province, China to find out the relationship between the coefficients of the equation and the geomorphic features based on VGE platform. It's exciting that the coefficient "a" appear to be correlative strongly with the stream scale, and the coefficient "b" may give an index to the valley shape. In the end, we use a sub-basin named Hao-Jia-Gou of the watershed as an application. The result of correlative research about quadratic regression equation and geomorphic features can save computing and improve the efficiency in silt dam systems planning.

Multisensor-based real-time quality monitoring by means of feature extraction, selection and modeling for Al alloy in arc welding

NASA Astrophysics Data System (ADS)

Zhang, Zhifen; Chen, Huabin; Xu, Yanling; Zhong, Jiyong; Lv, Na; Chen, Shanben

2015-08-01

Multisensory data fusion-based online welding quality monitoring has gained increasing attention in intelligent welding process. This paper mainly focuses on the automatic detection of typical welding defect for Al alloy in gas tungsten arc welding (GTAW) by means of analzing arc spectrum, sound and voltage signal. Based on the developed algorithms in time and frequency domain, 41 feature parameters were successively extracted from these signals to characterize the welding process and seam quality. Then, the proposed feature selection approach, i.e., hybrid fisher-based filter and wrapper was successfully utilized to evaluate the sensitivity of each feature and reduce the feature dimensions. Finally, the optimal feature subset with 19 features was selected to obtain the highest accuracy, i.e., 94.72% using established classification model. This study provides a guideline for feature extraction, selection and dynamic modeling based on heterogeneous multisensory data to achieve a reliable online defect detection system in arc welding.

Structured sparse linear graph embedding.

PubMed

Wang, Haixian

2012-03-01

Subspace learning is a core issue in pattern recognition and machine learning. Linear graph embedding (LGE) is a general framework for subspace learning. In this paper, we propose a structured sparse extension to LGE (SSLGE) by introducing a structured sparsity-inducing norm into LGE. Specifically, SSLGE casts the projection bases learning into a regression-type optimization problem, and then the structured sparsity regularization is applied to the regression coefficients. The regularization selects a subset of features and meanwhile encodes high-order information reflecting a priori structure information of the data. The SSLGE technique provides a unified framework for discovering structured sparse subspace. Computationally, by using a variational equality and the Procrustes transformation, SSLGE is efficiently solved with closed-form updates. Experimental results on face image show the effectiveness of the proposed method. Copyright © 2011 Elsevier Ltd. All rights reserved.

Classification of pre-sliced pork and Turkey ham qualities based on image colour and textural features and their relationships with consumer responses.

PubMed

Iqbal, Abdullah; Valous, Nektarios A; Mendoza, Fernando; Sun, Da-Wen; Allen, Paul

2010-03-01

Images of three qualities of pre-sliced pork and Turkey hams were evaluated for colour and textural features to characterize and classify them, and to model the ham appearance grading and preference responses of a group of consumers. A total of 26 colour features and 40 textural features were extracted for analysis. Using Mahalanobis distance and feature inter-correlation analyses, two best colour [mean of S (saturation in HSV colour space), std. deviation of b*, which indicates blue to yellow in L*a*b* colour space] and three textural features [entropy of b*, contrast of H (hue of HSV colour space), entropy of R (red of RGB colour space)] for pork, and three colour (mean of R, mean of H, std. deviation of a*, which indicates green to red in L*a*b* colour space) and two textural features [contrast of B, contrast of L* (luminance or lightness in L*a*b* colour space)] for Turkey hams were selected as features with the highest discriminant power. High classification performances were reached for both types of hams (>99.5% for pork and >90.5% for Turkey) using the best selected features or combinations of them. In spite of the poor/fair agreement among ham consumers as determined by Kappa analysis (Kappa-value<0.4) for sensory grading (surface colour, colour uniformity, bitonality, texture appearance and acceptability), a dichotomous logistic regression model using the best image features was able to explain the variability of consumers' responses for all sensorial attributes with accuracies higher than 74.1% for pork hams and 83.3% for Turkey hams. Copyright 2009 Elsevier Ltd. All rights reserved.

TU-CD-BRB-01: Normal Lung CT Texture Features Improve Predictive Models for Radiation Pneumonitis

DOE Office of Scientific and Technical Information (OSTI.GOV)

Krafft, S; The University of Texas Graduate School of Biomedical Sciences, Houston, TX; Briere, T

2015-06-15

Purpose: Existing normal tissue complication probability (NTCP) models for radiation pneumonitis (RP) traditionally rely on dosimetric and clinical data but are limited in terms of performance and generalizability. Extraction of pre-treatment image features provides a potential new category of data that can improve NTCP models for RP. We consider quantitative measures of total lung CT intensity and texture in a framework for prediction of RP. Methods: Available clinical and dosimetric data was collected for 198 NSCLC patients treated with definitive radiotherapy. Intensity- and texture-based image features were extracted from the T50 phase of the 4D-CT acquired for treatment planning. Amore » total of 3888 features (15 clinical, 175 dosimetric, and 3698 image features) were gathered and considered candidate predictors for modeling of RP grade≥3. A baseline logistic regression model with mean lung dose (MLD) was first considered. Additionally, a least absolute shrinkage and selection operator (LASSO) logistic regression was applied to the set of clinical and dosimetric features, and subsequently to the full set of clinical, dosimetric, and image features. Model performance was assessed by comparing area under the curve (AUC). Results: A simple logistic fit of MLD was an inadequate model of the data (AUC∼0.5). Including clinical and dosimetric parameters within the framework of the LASSO resulted in improved performance (AUC=0.648). Analysis of the full cohort of clinical, dosimetric, and image features provided further and significant improvement in model performance (AUC=0.727). Conclusions: To achieve significant gains in predictive modeling of RP, new categories of data should be considered in addition to clinical and dosimetric features. We have successfully incorporated CT image features into a framework for modeling RP and have demonstrated improved predictive performance. Validation and further investigation of CT image features in the context of RP NTCP modeling is warranted. This work was supported by the Rosalie B. Hite Fellowship in Cancer research awarded to SPK.« less

Enhancing the Discrimination Ability of a Gas Sensor Array Based on a Novel Feature Selection and Fusion Framework.

PubMed

Deng, Changjian; Lv, Kun; Shi, Debo; Yang, Bo; Yu, Song; He, Zhiyi; Yan, Jia

2018-06-12

In this paper, a novel feature selection and fusion framework is proposed to enhance the discrimination ability of gas sensor arrays for odor identification. Firstly, we put forward an efficient feature selection method based on the separability and the dissimilarity to determine the feature selection order for each type of feature when increasing the dimension of selected feature subsets. Secondly, the K-nearest neighbor (KNN) classifier is applied to determine the dimensions of the optimal feature subsets for different types of features. Finally, in the process of establishing features fusion, we come up with a classification dominance feature fusion strategy which conducts an effective basic feature. Experimental results on two datasets show that the recognition rates of Database I and Database II achieve 97.5% and 80.11%, respectively, when k = 1 for KNN classifier and the distance metric is correlation distance (COR), which demonstrates the superiority of the proposed feature selection and fusion framework in representing signal features. The novel feature selection method proposed in this paper can effectively select feature subsets that are conducive to the classification, while the feature fusion framework can fuse various features which describe the different characteristics of sensor signals, for enhancing the discrimination ability of gas sensors and, to a certain extent, suppressing drift effect.

Linking brain-wide multivoxel activation patterns to behaviour: Examples from language and math.

PubMed

Raizada, Rajeev D S; Tsao, Feng-Ming; Liu, Huei-Mei; Holloway, Ian D; Ansari, Daniel; Kuhl, Patricia K

2010-05-15

A key goal of cognitive neuroscience is to find simple and direct connections between brain and behaviour. However, fMRI analysis typically involves choices between many possible options, with each choice potentially biasing any brain-behaviour correlations that emerge. Standard methods of fMRI analysis assess each voxel individually, but then face the problem of selection bias when combining those voxels into a region-of-interest, or ROI. Multivariate pattern-based fMRI analysis methods use classifiers to analyse multiple voxels together, but can also introduce selection bias via data-reduction steps as feature selection of voxels, pre-selecting activated regions, or principal components analysis. We show here that strong brain-behaviour links can be revealed without any voxel selection or data reduction, using just plain linear regression as a classifier applied to the whole brain at once, i.e. treating each entire brain volume as a single multi-voxel pattern. The brain-behaviour correlations emerged despite the fact that the classifier was not provided with any information at all about subjects' behaviour, but instead was given only the neural data and its condition-labels. Surprisingly, more powerful classifiers such as a linear SVM and regularised logistic regression produce very similar results. We discuss some possible reasons why the very simple brain-wide linear regression model is able to find correlations with behaviour that are as strong as those obtained on the one hand from a specific ROI and on the other hand from more complex classifiers. In a manner which is unencumbered by arbitrary choices, our approach offers a method for investigating connections between brain and behaviour which is simple, rigorous and direct. Copyright (c) 2010 Elsevier Inc. All rights reserved.

Linking brain-wide multivoxel activation patterns to behaviour: Examples from language and math

PubMed Central

Raizada, Rajeev D.S.; Tsao, Feng-Ming; Liu, Huei-Mei; Holloway, Ian D.; Ansari, Daniel; Kuhl, Patricia K.

2010-01-01

A key goal of cognitive neuroscience is to find simple and direct connections between brain and behaviour. However, fMRI analysis typically involves choices between many possible options, with each choice potentially biasing any brain–behaviour correlations that emerge. Standard methods of fMRI analysis assess each voxel individually, but then face the problem of selection bias when combining those voxels into a region-of-interest, or ROI. Multivariate pattern-based fMRI analysis methods use classifiers to analyse multiple voxels together, but can also introduce selection bias via data-reduction steps as feature selection of voxels, pre-selecting activated regions, or principal components analysis. We show here that strong brain–behaviour links can be revealed without any voxel selection or data reduction, using just plain linear regression as a classifier applied to the whole brain at once, i.e. treating each entire brain volume as a single multi-voxel pattern. The brain–behaviour correlations emerged despite the fact that the classifier was not provided with any information at all about subjects' behaviour, but instead was given only the neural data and its condition-labels. Surprisingly, more powerful classifiers such as a linear SVM and regularised logistic regression produce very similar results. We discuss some possible reasons why the very simple brain-wide linear regression model is able to find correlations with behaviour that are as strong as those obtained on the one hand from a specific ROI and on the other hand from more complex classifiers. In a manner which is unencumbered by arbitrary choices, our approach offers a method for investigating connections between brain and behaviour which is simple, rigorous and direct. PMID:20132896

A web-based data visualization tool for the MIMIC-II database.

PubMed

Lee, Joon; Ribey, Evan; Wallace, James R

2016-02-04

Although MIMIC-II, a public intensive care database, has been recognized as an invaluable resource for many medical researchers worldwide, becoming a proficient MIMIC-II researcher requires knowledge of SQL programming and an understanding of the MIMIC-II database schema. These are challenging requirements especially for health researchers and clinicians who may have limited computer proficiency. In order to overcome this challenge, our objective was to create an interactive, web-based MIMIC-II data visualization tool that first-time MIMIC-II users can easily use to explore the database. The tool offers two main features: Explore and Compare. The Explore feature enables the user to select a patient cohort within MIMIC-II and visualize the distributions of various administrative, demographic, and clinical variables within the selected cohort. The Compare feature enables the user to select two patient cohorts and visually compare them with respect to a variety of variables. The tool is also helpful to experienced MIMIC-II researchers who can use it to substantially accelerate the cumbersome and time-consuming steps of writing SQL queries and manually visualizing extracted data. Any interested researcher can use the MIMIC-II data visualization tool for free to quickly and conveniently conduct a preliminary investigation on MIMIC-II with a few mouse clicks. Researchers can also use the tool to learn the characteristics of the MIMIC-II patients. Since it is still impossible to conduct multivariable regression inside the tool, future work includes adding analytics capabilities. Also, the next version of the tool will aim to utilize MIMIC-III which contains more data.

A combined Fisher and Laplacian score for feature selection in QSAR based drug design using compounds with known and unknown activities.

PubMed

Valizade Hasanloei, Mohammad Amin; Sheikhpour, Razieh; Sarram, Mehdi Agha; Sheikhpour, Elnaz; Sharifi, Hamdollah

2018-02-01

Quantitative structure-activity relationship (QSAR) is an effective computational technique for drug design that relates the chemical structures of compounds to their biological activities. Feature selection is an important step in QSAR based drug design to select the most relevant descriptors. One of the most popular feature selection methods for classification problems is Fisher score which aim is to minimize the within-class distance and maximize the between-class distance. In this study, the properties of Fisher criterion were extended for QSAR models to define the new distance metrics based on the continuous activity values of compounds with known activities. Then, a semi-supervised feature selection method was proposed based on the combination of Fisher and Laplacian criteria which exploits both compounds with known and unknown activities to select the relevant descriptors. To demonstrate the efficiency of the proposed semi-supervised feature selection method in selecting the relevant descriptors, we applied the method and other feature selection methods on three QSAR data sets such as serine/threonine-protein kinase PLK3 inhibitors, ROCK inhibitors and phenol compounds. The results demonstrated that the QSAR models built on the selected descriptors by the proposed semi-supervised method have better performance than other models. This indicates the efficiency of the proposed method in selecting the relevant descriptors using the compounds with known and unknown activities. The results of this study showed that the compounds with known and unknown activities can be helpful to improve the performance of the combined Fisher and Laplacian based feature selection methods.

A combined Fisher and Laplacian score for feature selection in QSAR based drug design using compounds with known and unknown activities

NASA Astrophysics Data System (ADS)

Valizade Hasanloei, Mohammad Amin; Sheikhpour, Razieh; Sarram, Mehdi Agha; Sheikhpour, Elnaz; Sharifi, Hamdollah

2018-02-01

Quantitative structure-activity relationship (QSAR) is an effective computational technique for drug design that relates the chemical structures of compounds to their biological activities. Feature selection is an important step in QSAR based drug design to select the most relevant descriptors. One of the most popular feature selection methods for classification problems is Fisher score which aim is to minimize the within-class distance and maximize the between-class distance. In this study, the properties of Fisher criterion were extended for QSAR models to define the new distance metrics based on the continuous activity values of compounds with known activities. Then, a semi-supervised feature selection method was proposed based on the combination of Fisher and Laplacian criteria which exploits both compounds with known and unknown activities to select the relevant descriptors. To demonstrate the efficiency of the proposed semi-supervised feature selection method in selecting the relevant descriptors, we applied the method and other feature selection methods on three QSAR data sets such as serine/threonine-protein kinase PLK3 inhibitors, ROCK inhibitors and phenol compounds. The results demonstrated that the QSAR models built on the selected descriptors by the proposed semi-supervised method have better performance than other models. This indicates the efficiency of the proposed method in selecting the relevant descriptors using the compounds with known and unknown activities. The results of this study showed that the compounds with known and unknown activities can be helpful to improve the performance of the combined Fisher and Laplacian based feature selection methods.

Comparisons and Selections of Features and Classifiers for Short Text Classification

NASA Astrophysics Data System (ADS)

Wang, Ye; Zhou, Zhi; Jin, Shan; Liu, Debin; Lu, Mi

2017-10-01

Short text is considerably different from traditional long text documents due to its shortness and conciseness, which somehow hinders the applications of conventional machine learning and data mining algorithms in short text classification. According to traditional artificial intelligence methods, we divide short text classification into three steps, namely preprocessing, feature selection and classifier comparison. In this paper, we have illustrated step-by-step how we approach our goals. Specifically, in feature selection, we compared the performance and robustness of the four methods of one-hot encoding, tf-idf weighting, word2vec and paragraph2vec, and in the classification part, we deliberately chose and compared Naive Bayes, Logistic Regression, Support Vector Machine, K-nearest Neighbor and Decision Tree as our classifiers. Then, we compared and analysed the classifiers horizontally with each other and vertically with feature selections. Regarding the datasets, we crawled more than 400,000 short text files from Shanghai and Shenzhen Stock Exchanges and manually labeled them into two classes, the big and the small. There are eight labels in the big class, and 59 labels in the small class.

A PCA aided cross-covariance scheme for discriminative feature extraction from EEG signals.

PubMed

Zarei, Roozbeh; He, Jing; Siuly, Siuly; Zhang, Yanchun

2017-07-01

Feature extraction of EEG signals plays a significant role in Brain-computer interface (BCI) as it can significantly affect the performance and the computational time of the system. The main aim of the current work is to introduce an innovative algorithm for acquiring reliable discriminating features from EEG signals to improve classification performances and to reduce the time complexity. This study develops a robust feature extraction method combining the principal component analysis (PCA) and the cross-covariance technique (CCOV) for the extraction of discriminatory information from the mental states based on EEG signals in BCI applications. We apply the correlation based variable selection method with the best first search on the extracted features to identify the best feature set for characterizing the distribution of mental state signals. To verify the robustness of the proposed feature extraction method, three machine learning techniques: multilayer perceptron neural networks (MLP), least square support vector machine (LS-SVM), and logistic regression (LR) are employed on the obtained features. The proposed methods are evaluated on two publicly available datasets. Furthermore, we evaluate the performance of the proposed methods by comparing it with some recently reported algorithms. The experimental results show that all three classifiers achieve high performance (above 99% overall classification accuracy) for the proposed feature set. Among these classifiers, the MLP and LS-SVM methods yield the best performance for the obtained feature. The average sensitivity, specificity and classification accuracy for these two classifiers are same, which are 99.32%, 100%, and 99.66%, respectively for the BCI competition dataset IVa and 100%, 100%, and 100%, for the BCI competition dataset IVb. The results also indicate the proposed methods outperform the most recently reported methods by at least 0.25% average accuracy improvement in dataset IVa. The execution time results show that the proposed method has less time complexity after feature selection. The proposed feature extraction method is very effective for getting representatives information from mental states EEG signals in BCI applications and reducing the computational complexity of classifiers by reducing the number of extracted features. Copyright © 2017 Elsevier B.V. All rights reserved.

Building a computer program to support children, parents, and distraction during healthcare procedures.

PubMed

Hanrahan, Kirsten; McCarthy, Ann Marie; Kleiber, Charmaine; Ataman, Kaan; Street, W Nick; Zimmerman, M Bridget; Ersig, Anne L

2012-10-01

This secondary data analysis used data mining methods to develop predictive models of child risk for distress during a healthcare procedure. Data used came from a study that predicted factors associated with children's responses to an intravenous catheter insertion while parents provided distraction coaching. From the 255 items used in the primary study, 44 predictive items were identified through automatic feature selection and used to build support vector machine regression models. Models were validated using multiple cross-validation tests and by comparing variables identified as explanatory in the traditional versus support vector machine regression. Rule-based approaches were applied to the model outputs to identify overall risk for distress. A decision tree was then applied to evidence-based instructions for tailoring distraction to characteristics and preferences of the parent and child. The resulting decision support computer application, titled Children, Parents and Distraction, is being used in research. Future use will support practitioners in deciding the level and type of distraction intervention needed by a child undergoing a healthcare procedure.

Object-based selection from spatially-invariant representations: evidence from a feature-report task.

PubMed

Matsukura, Michi; Vecera, Shaun P

2011-02-01

Attention selects objects as well as locations. When attention selects an object's features, observers identify two features from a single object more accurately than two features from two different objects (object-based effect of attention; e.g., Duncan, Journal of Experimental Psychology: General, 113, 501-517, 1984). Several studies have demonstrated that object-based attention can operate at a late visual processing stage that is independent of objects' spatial information (Awh, Dhaliwal, Christensen, & Matsukura, Psychological Science, 12, 329-334, 2001; Matsukura & Vecera, Psychonomic Bulletin & Review, 16, 529-536, 2009; Vecera, Journal of Experimental Psychology: General, 126, 14-18, 1997; Vecera & Farah, Journal of Experimental Psychology: General, 123, 146-160, 1994). In the present study, we asked two questions regarding this late object-based selection mechanism. In Part I, we investigated how observers' foreknowledge of to-be-reported features allows attention to select objects, as opposed to individual features. Using a feature-report task, a significant object-based effect was observed when to-be-reported features were known in advance but not when this advance knowledge was absent. In Part II, we examined what drives attention to select objects rather than individual features in the absence of observers' foreknowledge of to-be-reported features. Results suggested that, when there was no opportunity for observers to direct their attention to objects that possess to-be-reported features at the time of stimulus presentation, these stimuli must retain strong perceptual cues to establish themselves as separate objects.

RRegrs: an R package for computer-aided model selection with multiple regression models.

PubMed

Tsiliki, Georgia; Munteanu, Cristian R; Seoane, Jose A; Fernandez-Lozano, Carlos; Sarimveis, Haralambos; Willighagen, Egon L

2015-01-01

Predictive regression models can be created with many different modelling approaches. Choices need to be made for data set splitting, cross-validation methods, specific regression parameters and best model criteria, as they all affect the accuracy and efficiency of the produced predictive models, and therefore, raising model reproducibility and comparison issues. Cheminformatics and bioinformatics are extensively using predictive modelling and exhibit a need for standardization of these methodologies in order to assist model selection and speed up the process of predictive model development. A tool accessible to all users, irrespectively of their statistical knowledge, would be valuable if it tests several simple and complex regression models and validation schemes, produce unified reports, and offer the option to be integrated into more extensive studies. Additionally, such methodology should be implemented as a free programming package, in order to be continuously adapted and redistributed by others. We propose an integrated framework for creating multiple regression models, called RRegrs. The tool offers the option of ten simple and complex regression methods combined with repeated 10-fold and leave-one-out cross-validation. Methods include Multiple Linear regression, Generalized Linear Model with Stepwise Feature Selection, Partial Least Squares regression, Lasso regression, and Support Vector Machines Recursive Feature Elimination. The new framework is an automated fully validated procedure which produces standardized reports to quickly oversee the impact of choices in modelling algorithms and assess the model and cross-validation results. The methodology was implemented as an open source R package, available at https://www.github.com/enanomapper/RRegrs, by reusing and extending on the caret package. The universality of the new methodology is demonstrated using five standard data sets from different scientific fields. Its efficiency in cheminformatics and QSAR modelling is shown with three use cases: proteomics data for surface-modified gold nanoparticles, nano-metal oxides descriptor data, and molecular descriptors for acute aquatic toxicity data. The results show that for all data sets RRegrs reports models with equal or better performance for both training and test sets than those reported in the original publications. Its good performance as well as its adaptability in terms of parameter optimization could make RRegrs a popular framework to assist the initial exploration of predictive models, and with that, the design of more comprehensive in silico screening applications.Graphical abstractRRegrs is a computer-aided model selection framework for R multiple regression models; this is a fully validated procedure with application to QSAR modelling.

Methods for identifying SNP interactions: a review on variations of Logic Regression, Random Forest and Bayesian logistic regression.

PubMed

Chen, Carla Chia-Ming; Schwender, Holger; Keith, Jonathan; Nunkesser, Robin; Mengersen, Kerrie; Macrossan, Paula

2011-01-01

Due to advancements in computational ability, enhanced technology and a reduction in the price of genotyping, more data are being generated for understanding genetic associations with diseases and disorders. However, with the availability of large data sets comes the inherent challenges of new methods of statistical analysis and modeling. Considering a complex phenotype may be the effect of a combination of multiple loci, various statistical methods have been developed for identifying genetic epistasis effects. Among these methods, logic regression (LR) is an intriguing approach incorporating tree-like structures. Various methods have built on the original LR to improve different aspects of the model. In this study, we review four variations of LR, namely Logic Feature Selection, Monte Carlo Logic Regression, Genetic Programming for Association Studies, and Modified Logic Regression-Gene Expression Programming, and investigate the performance of each method using simulated and real genotype data. We contrast these with another tree-like approach, namely Random Forests, and a Bayesian logistic regression with stochastic search variable selection.

A Permutation Approach for Selecting the Penalty Parameter in Penalized Model Selection

PubMed Central

Sabourin, Jeremy A; Valdar, William; Nobel, Andrew B

2015-01-01

Summary We describe a simple, computationally effcient, permutation-based procedure for selecting the penalty parameter in LASSO penalized regression. The procedure, permutation selection, is intended for applications where variable selection is the primary focus, and can be applied in a variety of structural settings, including that of generalized linear models. We briefly discuss connections between permutation selection and existing theory for the LASSO. In addition, we present a simulation study and an analysis of real biomedical data sets in which permutation selection is compared with selection based on the following: cross-validation (CV), the Bayesian information criterion (BIC), Scaled Sparse Linear Regression, and a selection method based on recently developed testing procedures for the LASSO. PMID:26243050

DOE Office of Scientific and Technical Information (OSTI.GOV)

Rios Velazquez, E; Parmar, C; Narayan, V

Purpose: To compare the complementary value of quantitative radiomic features to that of radiologist-annotated semantic features in predicting EGFR mutations in lung adenocarcinomas. Methods: Pre-operative CT images of 258 lung adenocarcinoma patients were available. Tumors were segmented using the sing-click ensemble segmentation algorithm. A set of radiomic features was extracted using 3D-Slicer. Test-retest reproducibility and unsupervised dimensionality reduction were applied to select a subset of reproducible and independent radiomic features. Twenty semantic annotations were scored by an expert radiologist, describing the tumor, surrounding tissue and associated findings. Minimum-redundancy-maximum-relevance (MRMR) was used to identify the most informative radiomic and semantic featuresmore » in 172 patients (training-set, temporal split). Radiomic, semantic and combined radiomic-semantic logistic regression models to predict EGFR mutations were evaluated in and independent validation dataset of 86 patients using the area under the receiver operating curve (AUC). Results: EGFR mutations were found in 77/172 (45%) and 39/86 (45%) of the training and validation sets, respectively. Univariate AUCs showed a similar range for both feature types: radiomics median AUC = 0.57 (range: 0.50 – 0.62); semantic median AUC = 0.53 (range: 0.50 – 0.64, Wilcoxon p = 0.55). After MRMR feature selection, the best-performing radiomic, semantic, and radiomic-semantic logistic regression models, for EGFR mutations, showed a validation AUC of 0.56 (p = 0.29), 0.63 (p = 0.063) and 0.67 (p = 0.004), respectively. Conclusion: Quantitative volumetric and textural Radiomic features complement the qualitative and semi-quantitative radiologist annotations. The prognostic value of informative qualitative semantic features such as cavitation and lobulation is increased with the addition of quantitative textural features from the tumor region.« less

Relevance popularity: A term event model based feature selection scheme for text classification.

PubMed

Feng, Guozhong; An, Baiguo; Yang, Fengqin; Wang, Han; Zhang, Libiao

2017-01-01

Feature selection is a practical approach for improving the performance of text classification methods by optimizing the feature subsets input to classifiers. In traditional feature selection methods such as information gain and chi-square, the number of documents that contain a particular term (i.e. the document frequency) is often used. However, the frequency of a given term appearing in each document has not been fully investigated, even though it is a promising feature to produce accurate classifications. In this paper, we propose a new feature selection scheme based on a term event Multinomial naive Bayes probabilistic model. According to the model assumptions, the matching score function, which is based on the prediction probability ratio, can be factorized. Finally, we derive a feature selection measurement for each term after replacing inner parameters by their estimators. On a benchmark English text datasets (20 Newsgroups) and a Chinese text dataset (MPH-20), our numerical experiment results obtained from using two widely used text classifiers (naive Bayes and support vector machine) demonstrate that our method outperformed the representative feature selection methods.

An improved survivability prognosis of breast cancer by using sampling and feature selection technique to solve imbalanced patient classification data.

PubMed

Wang, Kung-Jeng; Makond, Bunjira; Wang, Kung-Min

2013-11-09

Breast cancer is one of the most critical cancers and is a major cause of cancer death among women. It is essential to know the survivability of the patients in order to ease the decision making process regarding medical treatment and financial preparation. Recently, the breast cancer data sets have been imbalanced (i.e., the number of survival patients outnumbers the number of non-survival patients) whereas the standard classifiers are not applicable for the imbalanced data sets. The methods to improve survivability prognosis of breast cancer need for study. Two well-known five-year prognosis models/classifiers [i.e., logistic regression (LR) and decision tree (DT)] are constructed by combining synthetic minority over-sampling technique (SMOTE), cost-sensitive classifier technique (CSC), under-sampling, bagging, and boosting. The feature selection method is used to select relevant variables, while the pruning technique is applied to obtain low information-burden models. These methods are applied on data obtained from the Surveillance, Epidemiology, and End Results database. The improvements of survivability prognosis of breast cancer are investigated based on the experimental results. Experimental results confirm that the DT and LR models combined with SMOTE, CSC, and under-sampling generate higher predictive performance consecutively than the original ones. Most of the time, DT and LR models combined with SMOTE and CSC use less informative burden/features when a feature selection method and a pruning technique are applied. LR is found to have better statistical power than DT in predicting five-year survivability. CSC is superior to SMOTE, under-sampling, bagging, and boosting to improve the prognostic performance of DT and LR.

An improved survivability prognosis of breast cancer by using sampling and feature selection technique to solve imbalanced patient classification data

PubMed Central

2013-01-01

Background Breast cancer is one of the most critical cancers and is a major cause of cancer death among women. It is essential to know the survivability of the patients in order to ease the decision making process regarding medical treatment and financial preparation. Recently, the breast cancer data sets have been imbalanced (i.e., the number of survival patients outnumbers the number of non-survival patients) whereas the standard classifiers are not applicable for the imbalanced data sets. The methods to improve survivability prognosis of breast cancer need for study. Methods Two well-known five-year prognosis models/classifiers [i.e., logistic regression (LR) and decision tree (DT)] are constructed by combining synthetic minority over-sampling technique (SMOTE) ,cost-sensitive classifier technique (CSC), under-sampling, bagging, and boosting. The feature selection method is used to select relevant variables, while the pruning technique is applied to obtain low information-burden models. These methods are applied on data obtained from the Surveillance, Epidemiology, and End Results database. The improvements of survivability prognosis of breast cancer are investigated based on the experimental results. Results Experimental results confirm that the DT and LR models combined with SMOTE, CSC, and under-sampling generate higher predictive performance consecutively than the original ones. Most of the time, DT and LR models combined with SMOTE and CSC use less informative burden/features when a feature selection method and a pruning technique are applied. Conclusions LR is found to have better statistical power than DT in predicting five-year survivability. CSC is superior to SMOTE, under-sampling, bagging, and boosting to improve the prognostic performance of DT and LR. PMID:24207108

Systematic review of the influence of foraging habitat on red-cockaded woodpecker reproductive success.

DOE Office of Scientific and Technical Information (OSTI.GOV)

Garabedian, James E.

Relationships between foraging habitat and reproductive success provide compelling evidence of the contribution of specific vegetative features to foraging habitat quality, a potentially limiting factor for many animal populations. For example, foraging habitat quality likely will gain importance in the recovery of the threatened red-cockaded woodpecker Picoides borealis (RCW) in the USA as immediate nesting constraints are mitigated. Several researchers have characterized resource selection by foraging RCWs, but emerging research linking reproductive success (e.g. clutch size, nestling and fledgling production, and group size) and foraging habitat features has yet to be synthesized. Therefore, we reviewed peer-refereed scientific literature and technicalmore » resources (e.g. books, symposia proceedings, and technical reports) that examined RCW foraging ecology, foraging habitat, or demography to evaluate evidence for effects of the key foraging habitat features described in the species’ recovery plan on group reproductive success. Fitness-based habitat models suggest foraging habitat with low to intermediate pine Pinus spp. densities, presence of large and old pines, minimal midstory development, and herbaceous groundcover support more productive RCW groups. However, the relationships between some foraging habitat features and RCW reproductive success are not well supported by empirical data. In addition, few regression models account for > 30% of variation in reproductive success, and unstandardized multiple and simple linear regression coefficient estimates typically range from -0.100 to 0.100, suggesting ancillary variables and perhaps indirect mechanisms influence reproductive success. These findings suggest additional research is needed to address uncertainty in relationships between foraging habitat features and RCW reproductive success and in the mechanisms underlying those relationships.« less

Toward optimal feature and time segment selection by divergence method for EEG signals classification.

PubMed

Wang, Jie; Feng, Zuren; Lu, Na; Luo, Jing

2018-06-01

Feature selection plays an important role in the field of EEG signals based motor imagery pattern classification. It is a process that aims to select an optimal feature subset from the original set. Two significant advantages involved are: lowering the computational burden so as to speed up the learning procedure and removing redundant and irrelevant features so as to improve the classification performance. Therefore, feature selection is widely employed in the classification of EEG signals in practical brain-computer interface systems. In this paper, we present a novel statistical model to select the optimal feature subset based on the Kullback-Leibler divergence measure, and automatically select the optimal subject-specific time segment. The proposed method comprises four successive stages: a broad frequency band filtering and common spatial pattern enhancement as preprocessing, features extraction by autoregressive model and log-variance, the Kullback-Leibler divergence based optimal feature and time segment selection and linear discriminate analysis classification. More importantly, this paper provides a potential framework for combining other feature extraction models and classification algorithms with the proposed method for EEG signals classification. Experiments on single-trial EEG signals from two public competition datasets not only demonstrate that the proposed method is effective in selecting discriminative features and time segment, but also show that the proposed method yields relatively better classification results in comparison with other competitive methods. Copyright © 2018 Elsevier Ltd. All rights reserved.

Deep learning ensemble with asymptotic techniques for oscillometric blood pressure estimation.

PubMed

Lee, Soojeong; Chang, Joon-Hyuk

2017-11-01

This paper proposes a deep learning based ensemble regression estimator with asymptotic techniques, and offers a method that can decrease uncertainty for oscillometric blood pressure (BP) measurements using the bootstrap and Monte-Carlo approach. While the former is used to estimate SBP and DBP, the latter attempts to determine confidence intervals (CIs) for SBP and DBP based on oscillometric BP measurements. This work originally employs deep belief networks (DBN)-deep neural networks (DNN) to effectively estimate BPs based on oscillometric measurements. However, there are some inherent problems with these methods. First, it is not easy to determine the best DBN-DNN estimator, and worthy information might be omitted when selecting one DBN-DNN estimator and discarding the others. Additionally, our input feature vectors, obtained from only five measurements per subject, represent a very small sample size; this is a critical weakness when using the DBN-DNN technique and can cause overfitting or underfitting, depending on the structure of the algorithm. To address these problems, an ensemble with an asymptotic approach (based on combining the bootstrap with the DBN-DNN technique) is utilized to generate the pseudo features needed to estimate the SBP and DBP. In the first stage, the bootstrap-aggregation technique is used to create ensemble parameters. Afterward, the AdaBoost approach is employed for the second-stage SBP and DBP estimation. We then use the bootstrap and Monte-Carlo techniques in order to determine the CIs based on the target BP estimated using the DBN-DNN ensemble regression estimator with the asymptotic technique in the third stage. The proposed method can mitigate the estimation uncertainty such as large the standard deviation of error (SDE) on comparing the proposed DBN-DNN ensemble regression estimator with the DBN-DNN single regression estimator, we identify that the SDEs of the SBP and DBP are reduced by 0.58 and 0.57  mmHg, respectively. These indicate that the proposed method actually enhances the performance by 9.18% and 10.88% compared with the DBN-DNN single estimator. The proposed methodology improves the accuracy of BP estimation and reduces the uncertainty for BP estimation. Copyright © 2017 Elsevier B.V. All rights reserved.

IPF-LASSO: Integrative L 1-Penalized Regression with Penalty Factors for Prediction Based on Multi-Omics Data

PubMed Central

Jiang, Xiaoyu; Fuchs, Mathias

2017-01-01

As modern biotechnologies advance, it has become increasingly frequent that different modalities of high-dimensional molecular data (termed “omics” data in this paper), such as gene expression, methylation, and copy number, are collected from the same patient cohort to predict the clinical outcome. While prediction based on omics data has been widely studied in the last fifteen years, little has been done in the statistical literature on the integration of multiple omics modalities to select a subset of variables for prediction, which is a critical task in personalized medicine. In this paper, we propose a simple penalized regression method to address this problem by assigning different penalty factors to different data modalities for feature selection and prediction. The penalty factors can be chosen in a fully data-driven fashion by cross-validation or by taking practical considerations into account. In simulation studies, we compare the prediction performance of our approach, called IPF-LASSO (Integrative LASSO with Penalty Factors) and implemented in the R package ipflasso, with the standard LASSO and sparse group LASSO. The use of IPF-LASSO is also illustrated through applications to two real-life cancer datasets. All data and codes are available on the companion website to ensure reproducibility. PMID:28546826

Comparison of Feature Selection Techniques in Machine Learning for Anatomical Brain MRI in Dementia.

PubMed

Tohka, Jussi; Moradi, Elaheh; Huttunen, Heikki

2016-07-01

We present a comparative split-half resampling analysis of various data driven feature selection and classification methods for the whole brain voxel-based classification analysis of anatomical magnetic resonance images. We compared support vector machines (SVMs), with or without filter based feature selection, several embedded feature selection methods and stability selection. While comparisons of the accuracy of various classification methods have been reported previously, the variability of the out-of-training sample classification accuracy and the set of selected features due to independent training and test sets have not been previously addressed in a brain imaging context. We studied two classification problems: 1) Alzheimer's disease (AD) vs. normal control (NC) and 2) mild cognitive impairment (MCI) vs. NC classification. In AD vs. NC classification, the variability in the test accuracy due to the subject sample did not vary between different methods and exceeded the variability due to different classifiers. In MCI vs. NC classification, particularly with a large training set, embedded feature selection methods outperformed SVM-based ones with the difference in the test accuracy exceeding the test accuracy variability due to the subject sample. The filter and embedded methods produced divergent feature patterns for MCI vs. NC classification that suggests the utility of the embedded feature selection for this problem when linked with the good generalization performance. The stability of the feature sets was strongly correlated with the number of features selected, weakly correlated with the stability of classification accuracy, and uncorrelated with the average classification accuracy.

Application of Multi-task Lasso Regression in the Stellar Parametrization

NASA Astrophysics Data System (ADS)

Chang, L. N.; Zhang, P. A.

2015-01-01

The multi-task learning approaches have attracted the increasing attention in the fields of machine learning, computer vision, and artificial intelligence. By utilizing the correlations in tasks, learning multiple related tasks simultaneously is better than learning each task independently. An efficient multi-task Lasso (Least Absolute Shrinkage Selection and Operator) regression algorithm is proposed in this paper to estimate the physical parameters of stellar spectra. It not only makes different physical parameters share the common features, but also can effectively preserve their own peculiar features. Experiments were done based on the ELODIE data simulated with the stellar atmospheric simulation model, and on the SDSS data released by the American large survey Sloan. The precision of the model is better than those of the methods in the related literature, especially for the acceleration of gravity (lg g) and the chemical abundance ([Fe/H]). In the experiments, we changed the resolution of the spectrum, and applied the noises with different signal-to-noise ratio (SNR) to the spectrum, so as to illustrate the stability of the model. The results show that the model is influenced by both the resolution and the noise. But the influence of the noise is larger than that of the resolution. In general, the multi-task Lasso regression algorithm is easy to operate, has a strong stability, and also can improve the overall accuracy of the model.

Revealing metabolite biomarkers for acupuncture treatment by linear programming based feature selection.

PubMed

Wang, Yong; Wu, Qiao-Feng; Chen, Chen; Wu, Ling-Yun; Yan, Xian-Zhong; Yu, Shu-Guang; Zhang, Xiang-Sun; Liang, Fan-Rong

2012-01-01

Acupuncture has been practiced in China for thousands of years as part of the Traditional Chinese Medicine (TCM) and has gradually accepted in western countries as an alternative or complementary treatment. However, the underlying mechanism of acupuncture, especially whether there exists any difference between varies acupoints, remains largely unknown, which hinders its widespread use. In this study, we develop a novel Linear Programming based Feature Selection method (LPFS) to understand the mechanism of acupuncture effect, at molecular level, by revealing the metabolite biomarkers for acupuncture treatment. Specifically, we generate and investigate the high-throughput metabolic profiles of acupuncture treatment at several acupoints in human. To select the subsets of metabolites that best characterize the acupuncture effect for each meridian point, an optimization model is proposed to identify biomarkers from high-dimensional metabolic data from case and control samples. Importantly, we use nearest centroid as the prototype to simultaneously minimize the number of selected features and the leave-one-out cross validation error of classifier. We compared the performance of LPFS to several state-of-the-art methods, such as SVM recursive feature elimination (SVM-RFE) and sparse multinomial logistic regression approach (SMLR). We find that our LPFS method tends to reveal a small set of metabolites with small standard deviation and large shifts, which exactly serves our requirement for good biomarker. Biologically, several metabolite biomarkers for acupuncture treatment are revealed and serve as the candidates for further mechanism investigation. Also biomakers derived from five meridian points, Zusanli (ST36), Liangmen (ST21), Juliao (ST3), Yanglingquan (GB34), and Weizhong (BL40), are compared for their similarity and difference, which provide evidence for the specificity of acupoints. Our result demonstrates that metabolic profiling might be a promising method to investigate the molecular mechanism of acupuncture. Comparing with other existing methods, LPFS shows better performance to select a small set of key molecules. In addition, LPFS is a general methodology and can be applied to other high-dimensional data analysis, for example cancer genomics.

Revealing metabolite biomarkers for acupuncture treatment by linear programming based feature selection

PubMed Central

2012-01-01

Background Acupuncture has been practiced in China for thousands of years as part of the Traditional Chinese Medicine (TCM) and has gradually accepted in western countries as an alternative or complementary treatment. However, the underlying mechanism of acupuncture, especially whether there exists any difference between varies acupoints, remains largely unknown, which hinders its widespread use. Results In this study, we develop a novel Linear Programming based Feature Selection method (LPFS) to understand the mechanism of acupuncture effect, at molecular level, by revealing the metabolite biomarkers for acupuncture treatment. Specifically, we generate and investigate the high-throughput metabolic profiles of acupuncture treatment at several acupoints in human. To select the subsets of metabolites that best characterize the acupuncture effect for each meridian point, an optimization model is proposed to identify biomarkers from high-dimensional metabolic data from case and control samples. Importantly, we use nearest centroid as the prototype to simultaneously minimize the number of selected features and the leave-one-out cross validation error of classifier. We compared the performance of LPFS to several state-of-the-art methods, such as SVM recursive feature elimination (SVM-RFE) and sparse multinomial logistic regression approach (SMLR). We find that our LPFS method tends to reveal a small set of metabolites with small standard deviation and large shifts, which exactly serves our requirement for good biomarker. Biologically, several metabolite biomarkers for acupuncture treatment are revealed and serve as the candidates for further mechanism investigation. Also biomakers derived from five meridian points, Zusanli (ST36), Liangmen (ST21), Juliao (ST3), Yanglingquan (GB34), and Weizhong (BL40), are compared for their similarity and difference, which provide evidence for the specificity of acupoints. Conclusions Our result demonstrates that metabolic profiling might be a promising method to investigate the molecular mechanism of acupuncture. Comparing with other existing methods, LPFS shows better performance to select a small set of key molecules. In addition, LPFS is a general methodology and can be applied to other high-dimensional data analysis, for example cancer genomics. PMID:23046877

Mortality Prediction Model of Septic Shock Patients Based on Routinely Recorded Data

PubMed Central

Carrara, Marta; Baselli, Giuseppe; Ferrario, Manuela

2015-01-01

We studied the problem of mortality prediction in two datasets, the first composed of 23 septic shock patients and the second composed of 73 septic subjects selected from the public database MIMIC-II. For each patient we derived hemodynamic variables, laboratory results, and clinical information of the first 48 hours after shock onset and we performed univariate and multivariate analyses to predict mortality in the following 7 days. The results show interesting features that individually identify significant differences between survivors and nonsurvivors and features which gain importance only when considered together with the others in a multivariate regression model. This preliminary study on two small septic shock populations represents a novel contribution towards new personalized models for an integration of multiparameter patient information to improve critical care management of shock patients. PMID:26557154

Brain-computer interface signal processing at the Wadsworth Center: mu and sensorimotor beta rhythms.

PubMed

McFarland, Dennis J; Krusienski, Dean J; Wolpaw, Jonathan R

2006-01-01

The Wadsworth brain-computer interface (BCI), based on mu and beta sensorimotor rhythms, uses one- and two-dimensional cursor movement tasks and relies on user training. This is a real-time closed-loop system. Signal processing consists of channel selection, spatial filtering, and spectral analysis. Feature translation uses a regression approach and normalization. Adaptation occurs at several points in this process on the basis of different criteria and methods. It can use either feedforward (e.g., estimating the signal mean for normalization) or feedback control (e.g., estimating feature weights for the prediction equation). We view this process as the interaction between a dynamic user and a dynamic system that coadapt over time. Understanding the dynamics of this interaction and optimizing its performance represent a major challenge for BCI research.

Temporal Correlation Mechanisms and Their Role in Feature Selection: A Single-Unit Study in Primate Somatosensory Cortex

PubMed Central

Gomez-Ramirez, Manuel; Trzcinski, Natalie K.; Mihalas, Stefan; Niebur, Ernst

2014-01-01

Studies in vision show that attention enhances the firing rates of cells when it is directed towards their preferred stimulus feature. However, it is unknown whether other sensory systems employ this mechanism to mediate feature selection within their modalities. Moreover, whether feature-based attention modulates the correlated activity of a population is unclear. Indeed, temporal correlation codes such as spike-synchrony and spike-count correlations (rsc) are believed to play a role in stimulus selection by increasing the signal and reducing the noise in a population, respectively. Here, we investigate (1) whether feature-based attention biases the correlated activity between neurons when attention is directed towards their common preferred feature, (2) the interplay between spike-synchrony and rsc during feature selection, and (3) whether feature attention effects are common across the visual and tactile systems. Single-unit recordings were made in secondary somatosensory cortex of three non-human primates while animals engaged in tactile feature (orientation and frequency) and visual discrimination tasks. We found that both firing rate and spike-synchrony between neurons with similar feature selectivity were enhanced when attention was directed towards their preferred feature. However, attention effects on spike-synchrony were twice as large as those on firing rate, and had a tighter relationship with behavioral performance. Further, we observed increased rsc when attention was directed towards the visual modality (i.e., away from touch). These data suggest that similar feature selection mechanisms are employed in vision and touch, and that temporal correlation codes such as spike-synchrony play a role in mediating feature selection. We posit that feature-based selection operates by implementing multiple mechanisms that reduce the overall noise levels in the neural population and synchronize activity across subpopulations that encode the relevant features of sensory stimuli. PMID:25423284

Genetic improvement in mastitis resistance: comparison of selection criteria from cross-sectional and random regression sire models for somatic cell score.

PubMed

Odegård, J; Klemetsdal, G; Heringstad, B

2005-04-01

Several selection criteria for reducing incidence of mastitis were developed from a random regression sire model for test-day somatic cell score (SCS). For comparison, sire transmitting abilities were also predicted based on a cross-sectional model for lactation mean SCS. Only first-crop daughters were used in genetic evaluation of SCS, and the different selection criteria were compared based on their correlation with incidence of clinical mastitis in second-crop daughters (measured as mean daughter deviations). Selection criteria were predicted based on both complete and reduced first-crop daughter groups (261 or 65 daughters per sire, respectively). For complete daughter groups, predicted transmitting abilities at around 30 d in milk showed the best predictive ability for incidence of clinical mastitis, closely followed by average predicted transmitting abilities over the entire lactation. Both of these criteria were derived from the random regression model. These selection criteria improved accuracy of selection by approximately 2% relative to a cross-sectional model. However, for reduced daughter groups, the cross-sectional model yielded increased predictive ability compared with the selection criteria based on the random regression model. This result may be explained by the cross-sectional model being more robust, i.e., less sensitive to precision of (co)variance components estimates and effects of data structure.

Feature selection methods for object-based classification of sub-decimeter resolution digital aerial imagery

USDA-ARS?s Scientific Manuscript database

Due to the availability of numerous spectral, spatial, and contextual features, the determination of optimal features and class separabilities can be a time consuming process in object-based image analysis (OBIA). While several feature selection methods have been developed to assist OBIA, a robust c...

Aesthetic perception of visual textures: a holistic exploration using texture analysis, psychological experiment, and perception modeling.

PubMed

Liu, Jianli; Lughofer, Edwin; Zeng, Xianyi

2015-01-01

Modeling human aesthetic perception of visual textures is important and valuable in numerous industrial domains, such as product design, architectural design, and decoration. Based on results from a semantic differential rating experiment, we modeled the relationship between low-level basic texture features and aesthetic properties involved in human aesthetic texture perception. First, we compute basic texture features from textural images using four classical methods. These features are neutral, objective, and independent of the socio-cultural context of the visual textures. Then, we conduct a semantic differential rating experiment to collect from evaluators their aesthetic perceptions of selected textural stimuli. In semantic differential rating experiment, eights pairs of aesthetic properties are chosen, which are strongly related to the socio-cultural context of the selected textures and to human emotions. They are easily understood and connected to everyday life. We propose a hierarchical feed-forward layer model of aesthetic texture perception and assign 8 pairs of aesthetic properties to different layers. Finally, we describe the generation of multiple linear and non-linear regression models for aesthetic prediction by taking dimensionality-reduced texture features and aesthetic properties of visual textures as dependent and independent variables, respectively. Our experimental results indicate that the relationships between each layer and its neighbors in the hierarchical feed-forward layer model of aesthetic texture perception can be fitted well by linear functions, and the models thus generated can successfully bridge the gap between computational texture features and aesthetic texture properties.

Constraint programming based biomarker optimization.

PubMed

Zhou, Manli; Luo, Youxi; Sun, Guoquan; Mai, Guoqin; Zhou, Fengfeng

2015-01-01

Efficient and intuitive characterization of biological big data is becoming a major challenge for modern bio-OMIC based scientists. Interactive visualization and exploration of big data is proven to be one of the successful solutions. Most of the existing feature selection algorithms do not allow the interactive inputs from users in the optimizing process of feature selection. This study investigates this question as fixing a few user-input features in the finally selected feature subset and formulates these user-input features as constraints for a programming model. The proposed algorithm, fsCoP (feature selection based on constrained programming), performs well similar to or much better than the existing feature selection algorithms, even with the constraints from both literature and the existing algorithms. An fsCoP biomarker may be intriguing for further wet lab validation, since it satisfies both the classification optimization function and the biomedical knowledge. fsCoP may also be used for the interactive exploration of bio-OMIC big data by interactively adding user-defined constraints for modeling.

Hyperparameterization of soil moisture statistical models for North America with Ensemble Learning Models (Elm)

NASA Astrophysics Data System (ADS)

Steinberg, P. D.; Brener, G.; Duffy, D.; Nearing, G. S.; Pelissier, C.

2017-12-01

Hyperparameterization, of statistical models, i.e. automated model scoring and selection, such as evolutionary algorithms, grid searches, and randomized searches, can improve forecast model skill by reducing errors associated with model parameterization, model structure, and statistical properties of training data. Ensemble Learning Models (Elm), and the related Earthio package, provide a flexible interface for automating the selection of parameters and model structure for machine learning models common in climate science and land cover classification, offering convenient tools for loading NetCDF, HDF, Grib, or GeoTiff files, decomposition methods like PCA and manifold learning, and parallel training and prediction with unsupervised and supervised classification, clustering, and regression estimators. Continuum Analytics is using Elm to experiment with statistical soil moisture forecasting based on meteorological forcing data from NASA's North American Land Data Assimilation System (NLDAS). There Elm is using the NSGA-2 multiobjective optimization algorithm for optimizing statistical preprocessing of forcing data to improve goodness-of-fit for statistical models (i.e. feature engineering). This presentation will discuss Elm and its components, including dask (distributed task scheduling), xarray (data structures for n-dimensional arrays), and scikit-learn (statistical preprocessing, clustering, classification, regression), and it will show how NSGA-2 is being used for automate selection of soil moisture forecast statistical models for North America.

A genetic programming approach to oral cancer prognosis

PubMed Central

Tan, Mei Sze; Tan, Jing Wei; Yap, Hwa Jen; Abdul Kareem, Sameem; Zain, Rosnah Binti

2016-01-01

Background The potential of genetic programming (GP) on various fields has been attained in recent years. In bio-medical field, many researches in GP are focused on the recognition of cancerous cells and also on gene expression profiling data. In this research, the aim is to study the performance of GP on the survival prediction of a small sample size of oral cancer prognosis dataset, which is the first study in the field of oral cancer prognosis. Method GP is applied on an oral cancer dataset that contains 31 cases collected from the Malaysia Oral Cancer Database and Tissue Bank System (MOCDTBS). The feature subsets that is automatically selected through GP were noted and the influences of this subset on the results of GP were recorded. In addition, a comparison between the GP performance and that of the Support Vector Machine (SVM) and logistic regression (LR) are also done in order to verify the predictive capabilities of the GP. Result The result shows that GP performed the best (average accuracy of 83.87% and average AUROC of 0.8341) when the features selected are smoking, drinking, chewing, histological differentiation of SCC, and oncogene p63. In addition, based on the comparison results, we found that the GP outperformed the SVM and LR in oral cancer prognosis. Discussion Some of the features in the dataset are found to be statistically co-related. This is because the accuracy of the GP prediction drops when one of the feature in the best feature subset is excluded. Thus, GP provides an automatic feature selection function, which chooses features that are highly correlated to the prognosis of oral cancer. This makes GP an ideal prediction model for cancer clinical and genomic data that can be used to aid physicians in their decision making stage of diagnosis or prognosis. PMID:27688975

Estimating the Biodegradability of Treated Sewage Samples Using Synchronous Fluorescence Spectra

PubMed Central

Lai, Tien M.; Shin, Jae-Ki; Hur, Jin

2011-01-01

Synchronous fluorescence spectra (SFS) and the first derivative spectra of the influent versus the effluent wastewater samples were compared and the use of fluorescence indices is suggested as a means to estimate the biodegradability of the effluent wastewater. Three distinct peaks were identified from the SFS of the effluent wastewater samples. Protein-like fluorescence (PLF) was reduced, whereas fulvic and/or humic-like fluorescence (HLF) were enhanced, suggesting that the two fluorescence characteristics may represent biodegradable and refractory components, respectively. Five fluorescence indices were selected for the biodegradability estimation based on the spectral features changing from the influent to the effluent. Among the selected indices, the relative distribution of PLF to the total fluorescence area of SFS (Index II) exhibited the highest correlation coefficient with total organic carbon (TOC)-based biodegradability, which was even higher than those obtained with the traditional oxygen demand-based parameters. A multiple regression analysis using Index II and the area ratio of PLF to HLF (Index III) demonstrated the enhancement of the correlations from 0.558 to 0.711 for TOC-based biodegradability. The multiple regression equation finally obtained was 0.148 × Index II − 4.964 × Index III − 0.001 and 0.046 × Index II − 1.128 × Index III + 0.026. The fluorescence indices proposed here are expected to be utilized for successful development of real-time monitoring using a simple fluorescence sensing device for the biodegradability of treated sewage. PMID:22164023

Estimating the biodegradability of treated sewage samples using synchronous fluorescence spectra.

PubMed

Lai, Tien M; Shin, Jae-Ki; Hur, Jin

2011-01-01

Synchronous fluorescence spectra (SFS) and the first derivative spectra of the influent versus the effluent wastewater samples were compared and the use of fluorescence indices is suggested as a means to estimate the biodegradability of the effluent wastewater. Three distinct peaks were identified from the SFS of the effluent wastewater samples. Protein-like fluorescence (PLF) was reduced, whereas fulvic and/or humic-like fluorescence (HLF) were enhanced, suggesting that the two fluorescence characteristics may represent biodegradable and refractory components, respectively. Five fluorescence indices were selected for the biodegradability estimation based on the spectral features changing from the influent to the effluent. Among the selected indices, the relative distribution of PLF to the total fluorescence area of SFS (Index II) exhibited the highest correlation coefficient with total organic carbon (TOC)-based biodegradability, which was even higher than those obtained with the traditional oxygen demand-based parameters. A multiple regression analysis using Index II and the area ratio of PLF to HLF (Index III) demonstrated the enhancement of the correlations from 0.558 to 0.711 for TOC-based biodegradability. The multiple regression equation finally obtained was 0.148 × Index II - 4.964 × Index III - 0.001 and 0.046 × Index II - 1.128 × Index III + 0.026. The fluorescence indices proposed here are expected to be utilized for successful development of real-time monitoring using a simple fluorescence sensing device for the biodegradability of treated sewage.

Error modeling for surrogates of dynamical systems using machine learning: Machine-learning-based error model for surrogates of dynamical systems

DOE PAGES

Trehan, Sumeet; Carlberg, Kevin T.; Durlofsky, Louis J.

2017-07-14

A machine learning–based framework for modeling the error introduced by surrogate models of parameterized dynamical systems is proposed. The framework entails the use of high-dimensional regression techniques (eg, random forests, and LASSO) to map a large set of inexpensively computed “error indicators” (ie, features) produced by the surrogate model at a given time instance to a prediction of the surrogate-model error in a quantity of interest (QoI). This eliminates the need for the user to hand-select a small number of informative features. The methodology requires a training set of parameter instances at which the time-dependent surrogate-model error is computed bymore » simulating both the high-fidelity and surrogate models. Using these training data, the method first determines regression-model locality (via classification or clustering) and subsequently constructs a “local” regression model to predict the time-instantaneous error within each identified region of feature space. We consider 2 uses for the resulting error model: (1) as a correction to the surrogate-model QoI prediction at each time instance and (2) as a way to statistically model arbitrary functions of the time-dependent surrogate-model error (eg, time-integrated errors). We then apply the proposed framework to model errors in reduced-order models of nonlinear oil-water subsurface flow simulations, with time-varying well-control (bottom-hole pressure) parameters. The reduced-order models used in this work entail application of trajectory piecewise linearization in conjunction with proper orthogonal decomposition. Moreover, when the first use of the method is considered, numerical experiments demonstrate consistent improvement in accuracy in the time-instantaneous QoI prediction relative to the original surrogate model, across a large number of test cases. When the second use is considered, results show that the proposed method provides accurate statistical predictions of the time- and well-averaged errors.« less

Error modeling for surrogates of dynamical systems using machine learning: Machine-learning-based error model for surrogates of dynamical systems

DOE Office of Scientific and Technical Information (OSTI.GOV)

Trehan, Sumeet; Carlberg, Kevin T.; Durlofsky, Louis J.

A machine learning–based framework for modeling the error introduced by surrogate models of parameterized dynamical systems is proposed. The framework entails the use of high-dimensional regression techniques (eg, random forests, and LASSO) to map a large set of inexpensively computed “error indicators” (ie, features) produced by the surrogate model at a given time instance to a prediction of the surrogate-model error in a quantity of interest (QoI). This eliminates the need for the user to hand-select a small number of informative features. The methodology requires a training set of parameter instances at which the time-dependent surrogate-model error is computed bymore » simulating both the high-fidelity and surrogate models. Using these training data, the method first determines regression-model locality (via classification or clustering) and subsequently constructs a “local” regression model to predict the time-instantaneous error within each identified region of feature space. We consider 2 uses for the resulting error model: (1) as a correction to the surrogate-model QoI prediction at each time instance and (2) as a way to statistically model arbitrary functions of the time-dependent surrogate-model error (eg, time-integrated errors). We then apply the proposed framework to model errors in reduced-order models of nonlinear oil-water subsurface flow simulations, with time-varying well-control (bottom-hole pressure) parameters. The reduced-order models used in this work entail application of trajectory piecewise linearization in conjunction with proper orthogonal decomposition. Moreover, when the first use of the method is considered, numerical experiments demonstrate consistent improvement in accuracy in the time-instantaneous QoI prediction relative to the original surrogate model, across a large number of test cases. When the second use is considered, results show that the proposed method provides accurate statistical predictions of the time- and well-averaged errors.« less

A Filter Feature Selection Method Based on MFA Score and Redundancy Excluding and It's Application to Tumor Gene Expression Data Analysis.

PubMed

Li, Jiangeng; Su, Lei; Pang, Zenan

2015-12-01

Feature selection techniques have been widely applied to tumor gene expression data analysis in recent years. A filter feature selection method named marginal Fisher analysis score (MFA score) which is based on graph embedding has been proposed, and it has been widely used mainly because it is superior to Fisher score. Considering the heavy redundancy in gene expression data, we proposed a new filter feature selection technique in this paper. It is named MFA score+ and is based on MFA score and redundancy excluding. We applied it to an artificial dataset and eight tumor gene expression datasets to select important features and then used support vector machine as the classifier to classify the samples. Compared with MFA score, t test and Fisher score, it achieved higher classification accuracy.

Super-resolution fusion of complementary panoramic images based on cross-selection kernel regression interpolation.

PubMed

Chen, Lidong; Basu, Anup; Zhang, Maojun; Wang, Wei; Liu, Yu

2014-03-20

A complementary catadioptric imaging technique was proposed to solve the problem of low and nonuniform resolution in omnidirectional imaging. To enhance this research, our paper focuses on how to generate a high-resolution panoramic image from the captured omnidirectional image. To avoid the interference between the inner and outer images while fusing the two complementary views, a cross-selection kernel regression method is proposed. First, in view of the complementarity of sampling resolution in the tangential and radial directions between the inner and the outer images, respectively, the horizontal gradients in the expected panoramic image are estimated based on the scattered neighboring pixels mapped from the outer, while the vertical gradients are estimated using the inner image. Then, the size and shape of the regression kernel are adaptively steered based on the local gradients. Furthermore, the neighboring pixels in the next interpolation step of kernel regression are also selected based on the comparison between the horizontal and vertical gradients. In simulation and real-image experiments, the proposed method outperforms existing kernel regression methods and our previous wavelet-based fusion method in terms of both visual quality and objective evaluation.

Habitat patch size and nesting success of yellow-breasted chats

Treesearch

Dick E. Burhans; Frank R. Thompson III

1999-01-01

We measured vegetation at shrub patches used for nesting by Yellow-breasted Chats (Icteria virens) to evaluate the importance of nesting habitat patch features on nest predation, cowbird parasitism, and nest site selection. Logistic regression models indicated that nests in small patches (average diameter

A comparison of three feature selection methods for object-based classification of sub-decimeter resolution UltraCam-L imagery

USDA-ARS?s Scientific Manuscript database

The availability of numerous spectral, spatial, and contextual features with object-based image analysis (OBIA) renders the selection of optimal features a time consuming and subjective process. While several feature election methods have been used in conjunction with OBIA, a robust comparison of th...

Multi-level gene/MiRNA feature selection using deep belief nets and active learning.

PubMed

Ibrahim, Rania; Yousri, Noha A; Ismail, Mohamed A; El-Makky, Nagwa M

2014-01-01

Selecting the most discriminative genes/miRNAs has been raised as an important task in bioinformatics to enhance disease classifiers and to mitigate the dimensionality curse problem. Original feature selection methods choose genes/miRNAs based on their individual features regardless of how they perform together. Considering group features instead of individual ones provides a better view for selecting the most informative genes/miRNAs. Recently, deep learning has proven its ability in representing the data in multiple levels of abstraction, allowing for better discrimination between different classes. However, the idea of using deep learning for feature selection is not widely used in the bioinformatics field yet. In this paper, a novel multi-level feature selection approach named MLFS is proposed for selecting genes/miRNAs based on expression profiles. The approach is based on both deep and active learning. Moreover, an extension to use the technique for miRNAs is presented by considering the biological relation between miRNAs and genes. Experimental results show that the approach was able to outperform classical feature selection methods in hepatocellular carcinoma (HCC) by 9%, lung cancer by 6% and breast cancer by around 10% in F1-measure. Results also show the enhancement in F1-measure of our approach over recently related work in [1] and [2].

Thermal stress characterization using the electro-mechanical impedance method

NASA Astrophysics Data System (ADS)

Zhu, Xuan; Lanza di Scalea, Francesco; Fateh, Mahmood

2017-04-01

This study examines the potential of the Electro-Mechanical Impedance (EMI) method to provide an estimation of the developed thermal stress in constrained bar-like structures. This non-invasive method features the easiness of implementation and interpretation, while it is notoriously known for being vulnerable to environmental variability. A comprehensive analytical model is proposed to relate the measured electric admittance signatures of the PZT element to temperature and uniaxial stress applied to the underlying structure. The model results compare favorably to the experimental ones, where the sensitivities of features extracted from the admittance signatures to the varying stress levels and temperatures are determined. Two temperature compensation frameworks are proposed to characterize the thermal stress states: (a) a regression model is established based on temperature-only tests, and the residuals from the thermal stress tests are then used to isolate the stress measurand; (b) the temperature-only tests are decomposed by Principle Components Analysis (PCA) and the feature vectors of the thermal stress tests are reconstructed after removal of the temperaturesensitive components. For both methods, the features were selected based on their performance in Receiver Operating Characteristic (ROC) curves. Experimental results on the Continuous Welded Rails (CWR) are shown to demonstrate the effectiveness of these temperature compensation methods.

Feature engineering for drug name recognition in biomedical texts: feature conjunction and feature selection.

PubMed

Liu, Shengyu; Tang, Buzhou; Chen, Qingcai; Wang, Xiaolong; Fan, Xiaoming

2015-01-01

Drug name recognition (DNR) is a critical step for drug information extraction. Machine learning-based methods have been widely used for DNR with various types of features such as part-of-speech, word shape, and dictionary feature. Features used in current machine learning-based methods are usually singleton features which may be due to explosive features and a large number of noisy features when singleton features are combined into conjunction features. However, singleton features that can only capture one linguistic characteristic of a word are not sufficient to describe the information for DNR when multiple characteristics should be considered. In this study, we explore feature conjunction and feature selection for DNR, which have never been reported. We intuitively select 8 types of singleton features and combine them into conjunction features in two ways. Then, Chi-square, mutual information, and information gain are used to mine effective features. Experimental results show that feature conjunction and feature selection can improve the performance of the DNR system with a moderate number of features and our DNR system significantly outperforms the best system in the DDIExtraction 2013 challenge.

Classification of motor imagery tasks for BCI with multiresolution analysis and multiobjective feature selection.

PubMed

Ortega, Julio; Asensio-Cubero, Javier; Gan, John Q; Ortiz, Andrés

2016-07-15

Brain-computer interfacing (BCI) applications based on the classification of electroencephalographic (EEG) signals require solving high-dimensional pattern classification problems with such a relatively small number of training patterns that curse of dimensionality problems usually arise. Multiresolution analysis (MRA) has useful properties for signal analysis in both temporal and spectral analysis, and has been broadly used in the BCI field. However, MRA usually increases the dimensionality of the input data. Therefore, some approaches to feature selection or feature dimensionality reduction should be considered for improving the performance of the MRA based BCI. This paper investigates feature selection in the MRA-based frameworks for BCI. Several wrapper approaches to evolutionary multiobjective feature selection are proposed with different structures of classifiers. They are evaluated by comparing with baseline methods using sparse representation of features or without feature selection. The statistical analysis, by applying the Kolmogorov-Smirnoff and Kruskal-Wallis tests to the means of the Kappa values evaluated by using the test patterns in each approach, has demonstrated some advantages of the proposed approaches. In comparison with the baseline MRA approach used in previous studies, the proposed evolutionary multiobjective feature selection approaches provide similar or even better classification performances, with significant reduction in the number of features that need to be computed.

Feature Selection for Speech Emotion Recognition in Spanish and Basque: On the Use of Machine Learning to Improve Human-Computer Interaction

PubMed Central

Arruti, Andoni; Cearreta, Idoia; Álvarez, Aitor; Lazkano, Elena; Sierra, Basilio

2014-01-01

Study of emotions in human–computer interaction is a growing research area. This paper shows an attempt to select the most significant features for emotion recognition in spoken Basque and Spanish Languages using different methods for feature selection. RekEmozio database was used as the experimental data set. Several Machine Learning paradigms were used for the emotion classification task. Experiments were executed in three phases, using different sets of features as classification variables in each phase. Moreover, feature subset selection was applied at each phase in order to seek for the most relevant feature subset. The three phases approach was selected to check the validity of the proposed approach. Achieved results show that an instance-based learning algorithm using feature subset selection techniques based on evolutionary algorithms is the best Machine Learning paradigm in automatic emotion recognition, with all different feature sets, obtaining a mean of 80,05% emotion recognition rate in Basque and a 74,82% in Spanish. In order to check the goodness of the proposed process, a greedy searching approach (FSS-Forward) has been applied and a comparison between them is provided. Based on achieved results, a set of most relevant non-speaker dependent features is proposed for both languages and new perspectives are suggested. PMID:25279686

Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses.

PubMed

Faul, Franz; Erdfelder, Edgar; Buchner, Axel; Lang, Albert-Georg

2009-11-01

G*Power is a free power analysis program for a variety of statistical tests. We present extensions and improvements of the version introduced by Faul, Erdfelder, Lang, and Buchner (2007) in the domain of correlation and regression analyses. In the new version, we have added procedures to analyze the power of tests based on (1) single-sample tetrachoric correlations, (2) comparisons of dependent correlations, (3) bivariate linear regression, (4) multiple linear regression based on the random predictor model, (5) logistic regression, and (6) Poisson regression. We describe these new features and provide a brief introduction to their scope and handling.

Adaptive runtime for a multiprocessing API

DOEpatents

Antao, Samuel F.; Bertolli, Carlo; Eichenberger, Alexandre E.; O'Brien, John K.

2016-11-15

A computer-implemented method includes selecting a runtime for executing a program. The runtime includes a first combination of feature implementations, where each feature implementation implements a feature of an application programming interface (API). Execution of the program is monitored, and the execution uses the runtime. Monitor data is generated based on the monitoring. A second combination of feature implementations are selected, by a computer processor, where the selection is based at least in part on the monitor data. The runtime is modified by activating the second combination of feature implementations to replace the first combination of feature implementations.

Adaptive runtime for a multiprocessing API

DOEpatents

Antao, Samuel F.; Bertolli, Carlo; Eichenberger, Alexandre E.; O'Brien, John K.

2016-10-11

A computer-implemented method includes selecting a runtime for executing a program. The runtime includes a first combination of feature implementations, where each feature implementation implements a feature of an application programming interface (API). Execution of the program is monitored, and the execution uses the runtime. Monitor data is generated based on the monitoring. A second combination of feature implementations are selected, by a computer processor, where the selection is based at least in part on the monitor data. The runtime is modified by activating the second combination of feature implementations to replace the first combination of feature implementations.

A new solar power output prediction based on hybrid forecast engine and decomposition model.

PubMed

Zhang, Weijiang; Dang, Hongshe; Simoes, Rolando

2018-06-12

Regarding to the growing trend of photovoltaic (PV) energy as a clean energy source in electrical networks and its uncertain nature, PV energy prediction has been proposed by researchers in recent decades. This problem is directly effects on operation in power network while, due to high volatility of this signal, an accurate prediction model is demanded. A new prediction model based on Hilbert Huang transform (HHT) and integration of improved empirical mode decomposition (IEMD) with feature selection and forecast engine is presented in this paper. The proposed approach is divided into three main sections. In the first section, the signal is decomposed by the proposed IEMD as an accurate decomposition tool. To increase the accuracy of the proposed method, a new interpolation method has been used instead of cubic spline curve (CSC) fitting in EMD. Then the obtained output is entered into the new feature selection procedure to choose the best candidate inputs. Finally, the signal is predicted by a hybrid forecast engine composed of support vector regression (SVR) based on an intelligent algorithm. The effectiveness of the proposed approach has been verified over a number of real-world engineering test cases in comparison with other well-known models. The obtained results prove the validity of the proposed method. Copyright © 2018 ISA. Published by Elsevier Ltd. All rights reserved.

Corona Enhancement and Mosaic Architecture for Prognosis and Selection Between of Liver Resection Versus Transcatheter Arterial Chemoembolization in Single Hepatocellular Carcinomas >5 cm Without Extrahepatic Metastases: An Imaging-Based Retrospective Study.

PubMed

Li, Meng; Xin, Yongjie; Fu, Sirui; Liu, Zaiyi; Li, Yong; Hu, Baoshan; Chen, Shuting; Liang, Changhong; Lu, Ligong

2016-01-01

Corona enhancement and mosaic architecture are 2 radiologic features of hepatocellular carcinoma (HCC). However, neither their prognostic values nor their impacts on the selection of liver resection (LR) versus transcatheter arterial chemoembolization (TACE) as treatment modalities have been established.We retrospectively analyzed 275 patients with a single HCC lesion >5 cm without extrahepatic metastasis treated with LR or TACE. In LR patients, the overall survival (OS) and time to progression (TTP) were compared between corona enhancement negative (corona-) versus positive (corona+) and mosaic architecture negative (mosaic-) versus positive (mosaic+) patients. Furthermore, by the combination of corona and mosaic, LR patients were divided into negative for both corona and mosaic patterns (LR-/-), positive for only 1 feature (LR+/-), and positive for both (LR+/+); their OS and TTP were compared to those of the TACE group. Cox regression was performed to identify independent factors for OS.In the survival plots for LR, corona- had better OS and TTP than corona+, and mosaic- had better OS than mosaic+. There was no significant difference in TTP between the subgroups. On Cox regression analysis, corona enhancement, but not mosaic architecture, was a significant factor for OS, whereas neither were a significant factor for TTP. In TACE patients, neither corona nor mosaic patterns had significant correlations with OS or TTP. In the whole population, LR-/ and LR+/- subgroups had similar OS, which was better than the LR+/+ and TACE groups. Moreover, LR-/- and LR+/- patients had better TTP than TACE patients, but there were no differences between the LR-/- versus LR+/-, LR-/ versus LR+/+, LR+/- versus LR+/+, and LR+/+ versus TACE groups. On Cox regression analysis, the presence of corona/mosaic patterns was an independent prognostic factor for OS.Our results showed that, for patients with a single HCC >5 cm without extrahepatic metastasis, corona and mosaic patterns are indicators of limited LR efficacy. When both of the features are present, TACE can be used instead of LR with no negative influence on survival.

Infrared face recognition based on LBP histogram and KW feature selection

NASA Astrophysics Data System (ADS)

Xie, Zhihua

2014-07-01

The conventional LBP-based feature as represented by the local binary pattern (LBP) histogram still has room for performance improvements. This paper focuses on the dimension reduction of LBP micro-patterns and proposes an improved infrared face recognition method based on LBP histogram representation. To extract the local robust features in infrared face images, LBP is chosen to get the composition of micro-patterns of sub-blocks. Based on statistical test theory, Kruskal-Wallis (KW) feature selection method is proposed to get the LBP patterns which are suitable for infrared face recognition. The experimental results show combination of LBP and KW features selection improves the performance of infrared face recognition, the proposed method outperforms the traditional methods based on LBP histogram, discrete cosine transform(DCT) or principal component analysis(PCA).

[Feature extraction for breast cancer data based on geometric algebra theory and feature selection using differential evolution].

PubMed

Li, Jing; Hong, Wenxue

2014-12-01

The feature extraction and feature selection are the important issues in pattern recognition. Based on the geometric algebra representation of vector, a new feature extraction method using blade coefficient of geometric algebra was proposed in this study. At the same time, an improved differential evolution (DE) feature selection method was proposed to solve the elevated high dimension issue. The simple linear discriminant analysis was used as the classifier. The result of the 10-fold cross-validation (10 CV) classification of public breast cancer biomedical dataset was more than 96% and proved superior to that of the original features and traditional feature extraction method.

Applying quantitative adiposity feature analysis models to predict benefit of bevacizumab-based chemotherapy in ovarian cancer patients

NASA Astrophysics Data System (ADS)

Wang, Yunzhi; Qiu, Yuchen; Thai, Theresa; More, Kathleen; Ding, Kai; Liu, Hong; Zheng, Bin

2016-03-01

How to rationally identify epithelial ovarian cancer (EOC) patients who will benefit from bevacizumab or other antiangiogenic therapies is a critical issue in EOC treatments. The motivation of this study is to quantitatively measure adiposity features from CT images and investigate the feasibility of predicting potential benefit of EOC patients with or without receiving bevacizumab-based chemotherapy treatment using multivariate statistical models built based on quantitative adiposity image features. A dataset involving CT images from 59 advanced EOC patients were included. Among them, 32 patients received maintenance bevacizumab after primary chemotherapy and the remaining 27 patients did not. We developed a computer-aided detection (CAD) scheme to automatically segment subcutaneous fat areas (VFA) and visceral fat areas (SFA) and then extracted 7 adiposity-related quantitative features. Three multivariate data analysis models (linear regression, logistic regression and Cox proportional hazards regression) were performed respectively to investigate the potential association between the model-generated prediction results and the patients' progression-free survival (PFS) and overall survival (OS). The results show that using all 3 statistical models, a statistically significant association was detected between the model-generated results and both of the two clinical outcomes in the group of patients receiving maintenance bevacizumab (p<0.01), while there were no significant association for both PFS and OS in the group of patients without receiving maintenance bevacizumab. Therefore, this study demonstrated the feasibility of using quantitative adiposity-related CT image features based statistical prediction models to generate a new clinical marker and predict the clinical outcome of EOC patients receiving maintenance bevacizumab-based chemotherapy.

New machine-learning algorithms for prediction of Parkinson's disease

NASA Astrophysics Data System (ADS)

Mandal, Indrajit; Sairam, N.

2014-03-01

This article presents an enhanced prediction accuracy of diagnosis of Parkinson's disease (PD) to prevent the delay and misdiagnosis of patients using the proposed robust inference system. New machine-learning methods are proposed and performance comparisons are based on specificity, sensitivity, accuracy and other measurable parameters. The robust methods of treating Parkinson's disease (PD) includes sparse multinomial logistic regression, rotation forest ensemble with support vector machines and principal components analysis, artificial neural networks, boosting methods. A new ensemble method comprising of the Bayesian network optimised by Tabu search algorithm as classifier and Haar wavelets as projection filter is used for relevant feature selection and ranking. The highest accuracy obtained by linear logistic regression and sparse multinomial logistic regression is 100% and sensitivity, specificity of 0.983 and 0.996, respectively. All the experiments are conducted over 95% and 99% confidence levels and establish the results with corrected t-tests. This work shows a high degree of advancement in software reliability and quality of the computer-aided diagnosis system and experimentally shows best results with supportive statistical inference.

Pharmacokinetic Tumor Heterogeneity as a Prognostic Biomarker for Classifying Breast Cancer Recurrence Risk.

PubMed

Mahrooghy, Majid; Ashraf, Ahmed B; Daye, Dania; McDonald, Elizabeth S; Rosen, Mark; Mies, Carolyn; Feldman, Michael; Kontos, Despina

2015-06-01

Heterogeneity in cancer can affect response to therapy and patient prognosis. Histologic measures have classically been used to measure heterogeneity, although a reliable noninvasive measurement is needed both to establish baseline risk of recurrence and monitor response to treatment. Here, we propose using spatiotemporal wavelet kinetic features from dynamic contrast-enhanced magnetic resonance imaging to quantify intratumor heterogeneity in breast cancer. Tumor pixels are first partitioned into homogeneous subregions using pharmacokinetic measures. Heterogeneity wavelet kinetic (HetWave) features are then extracted from these partitions to obtain spatiotemporal patterns of the wavelet coefficients and the contrast agent uptake. The HetWave features are evaluated in terms of their prognostic value using a logistic regression classifier with genetic algorithm wrapper-based feature selection to classify breast cancer recurrence risk as determined by a validated gene expression assay. Receiver operating characteristic analysis and area under the curve (AUC) are computed to assess classifier performance using leave-one-out cross validation. The HetWave features outperform other commonly used features (AUC = 0.88 HetWave versus 0.70 standard features). The combination of HetWave and standard features further increases classifier performance (AUCs 0.94). The rate of the spatial frequency pattern over the pharmacokinetic partitions can provide valuable prognostic information. HetWave could be a powerful feature extraction approach for characterizing tumor heterogeneity, providing valuable prognostic information.

AVC: Selecting discriminative features on basis of AUC by maximizing variable complementarity.

PubMed

Sun, Lei; Wang, Jun; Wei, Jinmao

2017-03-14

The Receiver Operator Characteristic (ROC) curve is well-known in evaluating classification performance in biomedical field. Owing to its superiority in dealing with imbalanced and cost-sensitive data, the ROC curve has been exploited as a popular metric to evaluate and find out disease-related genes (features). The existing ROC-based feature selection approaches are simple and effective in evaluating individual features. However, these approaches may fail to find real target feature subset due to their lack of effective means to reduce the redundancy between features, which is essential in machine learning. In this paper, we propose to assess feature complementarity by a trick of measuring the distances between the misclassified instances and their nearest misses on the dimensions of pairwise features. If a misclassified instance and its nearest miss on one feature dimension are far apart on another feature dimension, the two features are regarded as complementary to each other. Subsequently, we propose a novel filter feature selection approach on the basis of the ROC analysis. The new approach employs an efficient heuristic search strategy to select optimal features with highest complementarities. The experimental results on a broad range of microarray data sets validate that the classifiers built on the feature subset selected by our approach can get the minimal balanced error rate with a small amount of significant features. Compared with other ROC-based feature selection approaches, our new approach can select fewer features and effectively improve the classification performance.

Automatic learning-based beam angle selection for thoracic IMRT

DOE Office of Scientific and Technical Information (OSTI.GOV)

Amit, Guy; Marshall, Andrea; Purdie, Thomas G., E-mail: tom.purdie@rmp.uhn.ca

Purpose: The treatment of thoracic cancer using external beam radiation requires an optimal selection of the radiation beam directions to ensure effective coverage of the target volume and to avoid unnecessary treatment of normal healthy tissues. Intensity modulated radiation therapy (IMRT) planning is a lengthy process, which requires the planner to iterate between choosing beam angles, specifying dose–volume objectives and executing IMRT optimization. In thorax treatment planning, where there are no class solutions for beam placement, beam angle selection is performed manually, based on the planner’s clinical experience. The purpose of this work is to propose and study a computationallymore » efficient framework that utilizes machine learning to automatically select treatment beam angles. Such a framework may be helpful for reducing the overall planning workload. Methods: The authors introduce an automated beam selection method, based on learning the relationships between beam angles and anatomical features. Using a large set of clinically approved IMRT plans, a random forest regression algorithm is trained to map a multitude of anatomical features into an individual beam score. An optimization scheme is then built to select and adjust the beam angles, considering the learned interbeam dependencies. The validity and quality of the automatically selected beams evaluated using the manually selected beams from the corresponding clinical plans as the ground truth. Results: The analysis included 149 clinically approved thoracic IMRT plans. For a randomly selected test subset of 27 plans, IMRT plans were generated using automatically selected beams and compared to the clinical plans. The comparison of the predicted and the clinical beam angles demonstrated a good average correspondence between the two (angular distance 16.8° ± 10°, correlation 0.75 ± 0.2). The dose distributions of the semiautomatic and clinical plans were equivalent in terms of primary target volume coverage and organ at risk sparing and were superior over plans produced with fixed sets of common beam angles. The great majority of the automatic plans (93%) were approved as clinically acceptable by three radiation therapy specialists. Conclusions: The results demonstrated the feasibility of utilizing a learning-based approach for automatic selection of beam angles in thoracic IMRT planning. The proposed method may assist in reducing the manual planning workload, while sustaining plan quality.« less

Fragment and knowledge-based design of selective GSK-3beta inhibitors using virtual screening models.

PubMed

Vadivelan, S; Sinha, Barij Nayan; Tajne, Sunita; Jagarlapudi, Sarma A R P

2009-06-01

Glycogen Synthase Kinase 3beta is one of the important targets in the treatment of type II diabetes and Alzheimer's disease. Currently this target is in pursuit for type II diabetes and a few GSK-3beta inhibitors have been now advanced to Phases I and II of clinical trials. The best validated HypoGen model consists of four pharmacophore features; 1) two hydrogen bond acceptors, 2) one hydrogen bond donor and 3) one hydrophobic. This pharmacophore model correlates well with the docking model, one hydrogen bond acceptor is necessary for the H-bond interaction with VAL135, and second hydrogen bond acceptor is important for the H-bond interactions with ARG141 and the hydrophobic feature may be required for the weak H-bond interactions with ASP133. The comparative model was developed from analogue and structure-based models like Catalyst, Glide SP & XP, Gold Fitness & ChemScore and Ligand Fit using multiple linear regression analysis. A virtual library of 10,000 molecules was generated employing fragment and knowledge-based approach and the comparative model was used to predict the activities of these molecules. The H-bond with ARG141 appears to be unique to GSK-3beta and explains the high GSK-3beta selectivity observed for 1H-Quinazolin-4-ones and Benzo[e][1,3]oxazin-4-ones. This understanding of protein-ligand interactions and molecular recognition increases the rapid development of potent and selective inhibitors, and also helps to eliminate the increase in number of false positives and negatives.

Varied Rates of Implementation of Patient-Centered Medical Home Features and Residents' Perceptions of Their Importance Based on Practice Experience.

PubMed

Eiff, M Patrice; Green, Larry A; Jones, Geoff; Devlaeminck, Alex Verdieck; Waller, Elaine; Dexter, Eve; Marino, Miguel; Carney, Patricia A

2017-03-01

Little is known about how the patient-centered medical home (PCMH) is being implemented in residency practices. We describe both the trends in implementation of PCMH features and the influence that working with PCMH features has on resident attitudes toward their importance in 14 family medicine residencies associated with the P4 Project. We assessed 24 residency continuity clinics annually between 2007-2011 on presence or absence of PCMH features. Annual resident surveys (n=690) assessed perceptions of importance of PCMH features using a 4-point scale (not at all important to very important). We used generalized estimating equations logistic regression to assess trends and ordinal-response proportional odds regression models to determine if resident ratings of importance were associated with working with those features during training. Implementation of electronic health record (EHR) features increased significantly from 2007-2011, such as email communication with patients (33% to 67%), preventive services registries (23% to 64%), chronic disease registries (63% to 82%), and population-based quality assurance (46% to 79%). Team-based care was the only process of care feature to change significantly (54% to 93%). Residents with any exposure to EHR-based features had higher odds of rating the features more important compared to those with no exposure. We observed consistently lower odds of the resident rating process of care features as more important with any exposure compared to no exposure. Residencies engaged in educational transformation were more successful in implementing EHR-based PCMH features, and exposure during training appears to positively influence resident ratings of importance, while exposure to process of care features are slower to implement with less influence on importance ratings.

A feature selection approach towards progressive vector transmission over the Internet

NASA Astrophysics Data System (ADS)

Miao, Ru; Song, Jia; Feng, Min

2017-09-01

WebGIS has been applied for visualizing and sharing geospatial information popularly over the Internet. In order to improve the efficiency of the client applications, the web-based progressive vector transmission approach is proposed. Important features should be selected and transferred firstly, and the methods for measuring the importance of features should be further considered in the progressive transmission. However, studies on progressive transmission for large-volume vector data have mostly focused on map generalization in the field of cartography, but rarely discussed on the selection of geographic features quantitatively. This paper applies information theory for measuring the feature importance of vector maps. A measurement model for the amount of information of vector features is defined based upon the amount of information for dealing with feature selection issues. The measurement model involves geometry factor, spatial distribution factor and thematic attribute factor. Moreover, a real-time transport protocol (RTP)-based progressive transmission method is then presented to improve the transmission of vector data. To clearly demonstrate the essential methodology and key techniques, a prototype for web-based progressive vector transmission is presented, and an experiment of progressive selection and transmission for vector features is conducted. The experimental results indicate that our approach clearly improves the performance and end-user experience of delivering and manipulating large vector data over the Internet.

Applying a new unequally weighted feature fusion method to improve CAD performance of classifying breast lesions

NASA Astrophysics Data System (ADS)

Zargari Khuzani, Abolfazl; Danala, Gopichandh; Heidari, Morteza; Du, Yue; Mashhadi, Najmeh; Qiu, Yuchen; Zheng, Bin

2018-02-01

Higher recall rates are a major challenge in mammography screening. Thus, developing computer-aided diagnosis (CAD) scheme to classify between malignant and benign breast lesions can play an important role to improve efficacy of mammography screening. Objective of this study is to develop and test a unique image feature fusion framework to improve performance in classifying suspicious mass-like breast lesions depicting on mammograms. The image dataset consists of 302 suspicious masses detected on both craniocaudal and mediolateral-oblique view images. Amongst them, 151 were malignant and 151 were benign. The study consists of following 3 image processing and feature analysis steps. First, an adaptive region growing segmentation algorithm was used to automatically segment mass regions. Second, a set of 70 image features related to spatial and frequency characteristics of mass regions were initially computed. Third, a generalized linear regression model (GLM) based machine learning classifier combined with a bat optimization algorithm was used to optimally fuse the selected image features based on predefined assessment performance index. An area under ROC curve (AUC) with was used as a performance assessment index. Applying CAD scheme to the testing dataset, AUC was 0.75+/-0.04, which was significantly higher than using a single best feature (AUC=0.69+/-0.05) or the classifier with equally weighted features (AUC=0.73+/-0.05). This study demonstrated that comparing to the conventional equal-weighted approach, using an unequal-weighted feature fusion approach had potential to significantly improve accuracy in classifying between malignant and benign breast masses.

HIV-1 protease cleavage site prediction based on two-stage feature selection method.

PubMed

Niu, Bing; Yuan, Xiao-Cheng; Roeper, Preston; Su, Qiang; Peng, Chun-Rong; Yin, Jing-Yuan; Ding, Juan; Li, HaiPeng; Lu, Wen-Cong

2013-03-01

Knowledge of the mechanism of HIV protease cleavage specificity is critical to the design of specific and effective HIV inhibitors. Searching for an accurate, robust, and rapid method to correctly predict the cleavage sites in proteins is crucial when searching for possible HIV inhibitors. In this article, HIV-1 protease specificity was studied using the correlation-based feature subset (CfsSubset) selection method combined with Genetic Algorithms method. Thirty important biochemical features were found based on a jackknife test from the original data set containing 4,248 features. By using the AdaBoost method with the thirty selected features the prediction model yields an accuracy of 96.7% for the jackknife test and 92.1% for an independent set test, with increased accuracy over the original dataset by 6.7% and 77.4%, respectively. Our feature selection scheme could be a useful technique for finding effective competitive inhibitors of HIV protease.

Estimating Soil Moisture Using Polsar Data: a Machine Learning Approach

NASA Astrophysics Data System (ADS)

Khedri, E.; Hasanlou, M.; Tabatabaeenejad, A.

2017-09-01

Soil moisture is an important parameter that affects several environmental processes. This parameter has many important functions in numerous sciences including agriculture, hydrology, aerology, flood prediction, and drought occurrence. However, field procedures for moisture calculations are not feasible in a vast agricultural region territory. This is due to the difficulty in calculating soil moisture in vast territories and high-cost nature as well as spatial and local variability of soil moisture. Polarimetric synthetic aperture radar (PolSAR) imaging is a powerful tool for estimating soil moisture. These images provide a wide field of view and high spatial resolution. For estimating soil moisture, in this study, a model of support vector regression (SVR) is proposed based on obtained data from AIRSAR in 2003 in C, L, and P channels. In this endeavor, sequential forward selection (SFS) and sequential backward selection (SBS) are evaluated to select suitable features of polarized image dataset for high efficient modeling. We compare the obtained data with in-situ data. Output results show that the SBS-SVR method results in higher modeling accuracy compared to SFS-SVR model. Statistical parameters obtained from this method show an R2 of 97% and an RMSE of lower than 0.00041 (m3/m3) for P, L, and C channels, which has provided better accuracy compared to other feature selection algorithms.

Imaging-based biomarkers of cognitive performance in older adults constructed via high-dimensional pattern regression applied to MRI and PET.

PubMed

Wang, Ying; Goh, Joshua O; Resnick, Susan M; Davatzikos, Christos

2013-01-01

In this study, we used high-dimensional pattern regression methods based on structural (gray and white matter; GM and WM) and functional (positron emission tomography of regional cerebral blood flow; PET) brain data to identify cross-sectional imaging biomarkers of cognitive performance in cognitively normal older adults from the Baltimore Longitudinal Study of Aging (BLSA). We focused on specific components of executive and memory domains known to decline with aging, including manipulation, semantic retrieval, long-term memory (LTM), and short-term memory (STM). For each imaging modality, brain regions associated with each cognitive domain were generated by adaptive regional clustering. A relevance vector machine was adopted to model the nonlinear continuous relationship between brain regions and cognitive performance, with cross-validation to select the most informative brain regions (using recursive feature elimination) as imaging biomarkers and optimize model parameters. Predicted cognitive scores using our regression algorithm based on the resulting brain regions correlated well with actual performance. Also, regression models obtained using combined GM, WM, and PET imaging modalities outperformed models based on single modalities. Imaging biomarkers related to memory performance included the orbito-frontal and medial temporal cortical regions with LTM showing stronger correlation with the temporal lobe than STM. Brain regions predicting executive performance included orbito-frontal, and occipito-temporal areas. The PET modality had higher contribution to most cognitive domains except manipulation, which had higher WM contribution from the superior longitudinal fasciculus and the genu of the corpus callosum. These findings based on machine-learning methods demonstrate the importance of combining structural and functional imaging data in understanding complex cognitive mechanisms and also their potential usage as biomarkers that predict cognitive status.

Evaluating disease management programme effectiveness: an introduction to the regression discontinuity design.

PubMed

Linden, Ariel; Adams, John L; Roberts, Nancy

2006-04-01

Although disease management (DM) has been in existence for over a decade, there is still much uncertainty as to its effectiveness in improving health status and reducing medical cost. The main reason is that most programme evaluations typically follow weak observational study designs that are subject to bias, most notably selection bias and regression to the mean. The regression discontinuity (RD) design may be the best alternative to randomized studies for evaluating DM programme effectiveness. The most crucial element of the RD design is its use of a 'cut-off' score on a pre-test measure to determine assignment to intervention or control. A valuable feature of this technique is that the pre-test measure does not have to be the same as the outcome measure, thus maximizing the programme's ability to use research-based practice guidelines, survey instruments and other tools to identify those individuals in greatest need of the programme intervention. Similarly, the cut-off score can be based on clinical understanding of the disease process, empirically derived, or resource-based. In the RD design, programme effectiveness is determined by a change in the pre-post relationship at the cut-off point. While the RD design is uniquely suitable for DM programme evaluation, its success will depend, in large part, on fundamental changes being made in the way DM programmes identify and assign individuals to the programme intervention.

3D statistical shape models incorporating 3D random forest regression voting for robust CT liver segmentation

NASA Astrophysics Data System (ADS)

Norajitra, Tobias; Meinzer, Hans-Peter; Maier-Hein, Klaus H.

2015-03-01

During image segmentation, 3D Statistical Shape Models (SSM) usually conduct a limited search for target landmarks within one-dimensional search profiles perpendicular to the model surface. In addition, landmark appearance is modeled only locally based on linear profiles and weak learners, altogether leading to segmentation errors from landmark ambiguities and limited search coverage. We present a new method for 3D SSM segmentation based on 3D Random Forest Regression Voting. For each surface landmark, a Random Regression Forest is trained that learns a 3D spatial displacement function between the according reference landmark and a set of surrounding sample points, based on an infinite set of non-local randomized 3D Haar-like features. Landmark search is then conducted omni-directionally within 3D search spaces, where voxelwise forest predictions on landmark position contribute to a common voting map which reflects the overall position estimate. Segmentation experiments were conducted on a set of 45 CT volumes of the human liver, of which 40 images were randomly chosen for training and 5 for testing. Without parameter optimization, using a simple candidate selection and a single resolution approach, excellent results were achieved, while faster convergence and better concavity segmentation were observed, altogether underlining the potential of our approach in terms of increased robustness from distinct landmark detection and from better search coverage.

Stochastic model search with binary outcomes for genome-wide association studies.

PubMed

Russu, Alberto; Malovini, Alberto; Puca, Annibale A; Bellazzi, Riccardo

2012-06-01

The spread of case-control genome-wide association studies (GWASs) has stimulated the development of new variable selection methods and predictive models. We introduce a novel Bayesian model search algorithm, Binary Outcome Stochastic Search (BOSS), which addresses the model selection problem when the number of predictors far exceeds the number of binary responses. Our method is based on a latent variable model that links the observed outcomes to the underlying genetic variables. A Markov Chain Monte Carlo approach is used for model search and to evaluate the posterior probability of each predictor. BOSS is compared with three established methods (stepwise regression, logistic lasso, and elastic net) in a simulated benchmark. Two real case studies are also investigated: a GWAS on the genetic bases of longevity, and the type 2 diabetes study from the Wellcome Trust Case Control Consortium. Simulations show that BOSS achieves higher precisions than the reference methods while preserving good recall rates. In both experimental studies, BOSS successfully detects genetic polymorphisms previously reported to be associated with the analyzed phenotypes. BOSS outperforms the other methods in terms of F-measure on simulated data. In the two real studies, BOSS successfully detects biologically relevant features, some of which are missed by univariate analysis and the three reference techniques. The proposed algorithm is an advance in the methodology for model selection with a large number of features. Our simulated and experimental results showed that BOSS proves effective in detecting relevant markers while providing a parsimonious model.

Folded concave penalized learning in identifying multimodal MRI marker for Parkinson’s disease

PubMed Central

Liu, Hongcheng; Du, Guangwei; Zhang, Lijun; Lewis, Mechelle M.; Wang, Xue; Yao, Tao; Li, Runze; Huang, Xuemei

2016-01-01

Background Brain MRI holds promise to gauge different aspects of Parkinson’s disease (PD)-related pathological changes. Its analysis, however, is hindered by the high-dimensional nature of the data. New method This study introduces folded concave penalized (FCP) sparse logistic regression to identify biomarkers for PD from a large number of potential factors. The proposed statistical procedures target the challenges of high-dimensionality with limited data samples acquired. The maximization problem associated with the sparse logistic regression model is solved by local linear approximation. The proposed procedures then are applied to the empirical analysis of multimodal MRI data. Results From 45 features, the proposed approach identified 15 MRI markers and the UPSIT, which are known to be clinically relevant to PD. By combining the MRI and clinical markers, we can enhance substantially the specificity and sensitivity of the model, as indicated by the ROC curves. Comparison to existing methods We compare the folded concave penalized learning scheme with both the Lasso penalized scheme and the principle component analysis-based feature selection (PCA) in the Parkinson’s biomarker identification problem that takes into account both the clinical features and MRI markers. The folded concave penalty method demonstrates a substantially better clinical potential than both the Lasso and PCA in terms of specificity and sensitivity. Conclusions For the first time, we applied the FCP learning method to MRI biomarker discovery in PD. The proposed approach successfully identified MRI markers that are clinically relevant. Combining these biomarkers with clinical features can substantially enhance performance. PMID:27102045

A novel feature extraction approach for microarray data based on multi-algorithm fusion

PubMed Central

Jiang, Zhu; Xu, Rong

2015-01-01

Feature extraction is one of the most important and effective method to reduce dimension in data mining, with emerging of high dimensional data such as microarray gene expression data. Feature extraction for gene selection, mainly serves two purposes. One is to identify certain disease-related genes. The other is to find a compact set of discriminative genes to build a pattern classifier with reduced complexity and improved generalization capabilities. Depending on the purpose of gene selection, two types of feature extraction algorithms including ranking-based feature extraction and set-based feature extraction are employed in microarray gene expression data analysis. In ranking-based feature extraction, features are evaluated on an individual basis, without considering inter-relationship between features in general, while set-based feature extraction evaluates features based on their role in a feature set by taking into account dependency between features. Just as learning methods, feature extraction has a problem in its generalization ability, which is robustness. However, the issue of robustness is often overlooked in feature extraction. In order to improve the accuracy and robustness of feature extraction for microarray data, a novel approach based on multi-algorithm fusion is proposed. By fusing different types of feature extraction algorithms to select the feature from the samples set, the proposed approach is able to improve feature extraction performance. The new approach is tested against gene expression dataset including Colon cancer data, CNS data, DLBCL data, and Leukemia data. The testing results show that the performance of this algorithm is better than existing solutions. PMID:25780277

A novel feature extraction approach for microarray data based on multi-algorithm fusion.

PubMed

Jiang, Zhu; Xu, Rong

2015-01-01

Feature extraction is one of the most important and effective method to reduce dimension in data mining, with emerging of high dimensional data such as microarray gene expression data. Feature extraction for gene selection, mainly serves two purposes. One is to identify certain disease-related genes. The other is to find a compact set of discriminative genes to build a pattern classifier with reduced complexity and improved generalization capabilities. Depending on the purpose of gene selection, two types of feature extraction algorithms including ranking-based feature extraction and set-based feature extraction are employed in microarray gene expression data analysis. In ranking-based feature extraction, features are evaluated on an individual basis, without considering inter-relationship between features in general, while set-based feature extraction evaluates features based on their role in a feature set by taking into account dependency between features. Just as learning methods, feature extraction has a problem in its generalization ability, which is robustness. However, the issue of robustness is often overlooked in feature extraction. In order to improve the accuracy and robustness of feature extraction for microarray data, a novel approach based on multi-algorithm fusion is proposed. By fusing different types of feature extraction algorithms to select the feature from the samples set, the proposed approach is able to improve feature extraction performance. The new approach is tested against gene expression dataset including Colon cancer data, CNS data, DLBCL data, and Leukemia data. The testing results show that the performance of this algorithm is better than existing solutions.

Multi-scale Gaussian representation and outline-learning based cell image segmentation.

PubMed

Farhan, Muhammad; Ruusuvuori, Pekka; Emmenlauer, Mario; Rämö, Pauli; Dehio, Christoph; Yli-Harja, Olli

2013-01-01

High-throughput genome-wide screening to study gene-specific functions, e.g. for drug discovery, demands fast automated image analysis methods to assist in unraveling the full potential of such studies. Image segmentation is typically at the forefront of such analysis as the performance of the subsequent steps, for example, cell classification, cell tracking etc., often relies on the results of segmentation. We present a cell cytoplasm segmentation framework which first separates cell cytoplasm from image background using novel approach of image enhancement and coefficient of variation of multi-scale Gaussian scale-space representation. A novel outline-learning based classification method is developed using regularized logistic regression with embedded feature selection which classifies image pixels as outline/non-outline to give cytoplasm outlines. Refinement of the detected outlines to separate cells from each other is performed in a post-processing step where the nuclei segmentation is used as contextual information. We evaluate the proposed segmentation methodology using two challenging test cases, presenting images with completely different characteristics, with cells of varying size, shape, texture and degrees of overlap. The feature selection and classification framework for outline detection produces very simple sparse models which use only a small subset of the large, generic feature set, that is, only 7 and 5 features for the two cases. Quantitative comparison of the results for the two test cases against state-of-the-art methods show that our methodology outperforms them with an increase of 4-9% in segmentation accuracy with maximum accuracy of 93%. Finally, the results obtained for diverse datasets demonstrate that our framework not only produces accurate segmentation but also generalizes well to different segmentation tasks.

Multi-scale Gaussian representation and outline-learning based cell image segmentation

PubMed Central

2013-01-01

Background High-throughput genome-wide screening to study gene-specific functions, e.g. for drug discovery, demands fast automated image analysis methods to assist in unraveling the full potential of such studies. Image segmentation is typically at the forefront of such analysis as the performance of the subsequent steps, for example, cell classification, cell tracking etc., often relies on the results of segmentation. Methods We present a cell cytoplasm segmentation framework which first separates cell cytoplasm from image background using novel approach of image enhancement and coefficient of variation of multi-scale Gaussian scale-space representation. A novel outline-learning based classification method is developed using regularized logistic regression with embedded feature selection which classifies image pixels as outline/non-outline to give cytoplasm outlines. Refinement of the detected outlines to separate cells from each other is performed in a post-processing step where the nuclei segmentation is used as contextual information. Results and conclusions We evaluate the proposed segmentation methodology using two challenging test cases, presenting images with completely different characteristics, with cells of varying size, shape, texture and degrees of overlap. The feature selection and classification framework for outline detection produces very simple sparse models which use only a small subset of the large, generic feature set, that is, only 7 and 5 features for the two cases. Quantitative comparison of the results for the two test cases against state-of-the-art methods show that our methodology outperforms them with an increase of 4-9% in segmentation accuracy with maximum accuracy of 93%. Finally, the results obtained for diverse datasets demonstrate that our framework not only produces accurate segmentation but also generalizes well to different segmentation tasks. PMID:24267488

QSPR models for half-wave reduction potential of steroids: a comparative study between feature selection and feature extraction from subsets of or entire set of descriptors.

PubMed

Hemmateenejad, Bahram; Yazdani, Mahdieh

2009-02-16

Steroids are widely distributed in nature and are found in plants, animals, and fungi in abundance. A data set consists of a diverse set of steroids have been used to develop quantitative structure-electrochemistry relationship (QSER) models for their half-wave reduction potential. Modeling was established by means of multiple linear regression (MLR) and principle component regression (PCR) analyses. In MLR analysis, the QSPR models were constructed by first grouping descriptors and then stepwise selection of variables from each group (MLR1) and stepwise selection of predictor variables from the pool of all calculated descriptors (MLR2). Similar procedure was used in PCR analysis so that the principal components (or features) were extracted from different group of descriptors (PCR1) and from entire set of descriptors (PCR2). The resulted models were evaluated using cross-validation, chance correlation, application to prediction reduction potential of some test samples and accessing applicability domain. Both MLR approaches represented accurate results however the QSPR model found by MLR1 was statistically more significant. PCR1 approach produced a model as accurate as MLR approaches whereas less accurate results were obtained by PCR2 approach. In overall, the correlation coefficients of cross-validation and prediction of the QSPR models resulted from MLR1, MLR2 and PCR1 approaches were higher than 90%, which show the high ability of the models to predict reduction potential of the studied steroids.

Building a Computer Program to Support Children, Parents, and Distraction during Healthcare Procedures

PubMed Central

McCarthy, Ann Marie; Kleiber, Charmaine; Ataman, Kaan; Street, W. Nick; Zimmerman, M. Bridget; Ersig, Anne L.

2012-01-01

This secondary data analysis used data mining methods to develop predictive models of child risk for distress during a healthcare procedure. Data used came from a study that predicted factors associated with children’s responses to an intravenous catheter insertion while parents provided distraction coaching. From the 255 items used in the primary study, 44 predictive items were identified through automatic feature selection and used to build support vector machine regression models. Models were validated using multiple cross-validation tests and by comparing variables identified as explanatory in the traditional versus support vector machine regression. Rule-based approaches were applied to the model outputs to identify overall risk for distress. A decision tree was then applied to evidence-based instructions for tailoring distraction to characteristics and preferences of the parent and child. The resulting decision support computer application, the Children, Parents and Distraction (CPaD), is being used in research. Future use will support practitioners in deciding the level and type of distraction intervention needed by a child undergoing a healthcare procedure. PMID:22805121

A CNN Regression Approach for Real-Time 2D/3D Registration.

PubMed

Shun Miao; Wang, Z Jane; Rui Liao

2016-05-01

In this paper, we present a Convolutional Neural Network (CNN) regression approach to address the two major limitations of existing intensity-based 2-D/3-D registration technology: 1) slow computation and 2) small capture range. Different from optimization-based methods, which iteratively optimize the transformation parameters over a scalar-valued metric function representing the quality of the registration, the proposed method exploits the information embedded in the appearances of the digitally reconstructed radiograph and X-ray images, and employs CNN regressors to directly estimate the transformation parameters. An automatic feature extraction step is introduced to calculate 3-D pose-indexed features that are sensitive to the variables to be regressed while robust to other factors. The CNN regressors are then trained for local zones and applied in a hierarchical manner to break down the complex regression task into multiple simpler sub-tasks that can be learned separately. Weight sharing is furthermore employed in the CNN regression model to reduce the memory footprint. The proposed approach has been quantitatively evaluated on 3 potential clinical applications, demonstrating its significant advantage in providing highly accurate real-time 2-D/3-D registration with a significantly enlarged capture range when compared to intensity-based methods.

Application of Multi-task Lasso Regression in the Parametrization of Stellar Spectra

NASA Astrophysics Data System (ADS)

Chang, Li-Na; Zhang, Pei-Ai

2015-07-01

The multi-task learning approaches have attracted the increasing attention in the fields of machine learning, computer vision, and artificial intelligence. By utilizing the correlations in tasks, learning multiple related tasks simultaneously is better than learning each task independently. An efficient multi-task Lasso (Least Absolute Shrinkage Selection and Operator) regression algorithm is proposed in this paper to estimate the physical parameters of stellar spectra. It not only can obtain the information about the common features of the different physical parameters, but also can preserve effectively their own peculiar features. Experiments were done based on the ELODIE synthetic spectral data simulated with the stellar atmospheric model, and on the SDSS data released by the American large-scale survey Sloan. The estimation precision of our model is better than those of the methods in the related literature, especially for the estimates of the gravitational acceleration (lg g) and the chemical abundance ([Fe/H]). In the experiments we changed the spectral resolution, and applied the noises with different signal-to-noise ratios (SNRs) to the spectral data, so as to illustrate the stability of the model. The results show that the model is influenced by both the resolution and the noise. But the influence of the noise is larger than that of the resolution. In general, the multi-task Lasso regression algorithm is easy to operate, it has a strong stability, and can also improve the overall prediction accuracy of the model.

System Complexity Reduction via Feature Selection

ERIC Educational Resources Information Center

Deng, Houtao

2011-01-01

This dissertation transforms a set of system complexity reduction problems to feature selection problems. Three systems are considered: classification based on association rules, network structure learning, and time series classification. Furthermore, two variable importance measures are proposed to reduce the feature selection bias in tree…

Volatility forecasting for low-volatility portfolio selection in the US and the Korean equity markets

NASA Astrophysics Data System (ADS)

Kim, Saejoon

2018-01-01

We consider the problem of low-volatility portfolio selection which has been the subject of extensive research in the field of portfolio selection. To improve the currently existing techniques that rely purely on past information to select low-volatility portfolios, this paper investigates the use of time series regression techniques that make forecasts of future volatility to select the portfolios. In particular, for the first time, the utility of support vector regression and its enhancements as portfolio selection techniques is provided. It is shown that our regression-based portfolio selection provides attractive outperformances compared to the benchmark index and the portfolio defined by a well-known strategy on the data-sets of the S&P 500 and the KOSPI 200.

IMMAN: free software for information theory-based chemometric analysis.

PubMed

Urias, Ricardo W Pino; Barigye, Stephen J; Marrero-Ponce, Yovani; García-Jacas, César R; Valdes-Martiní, José R; Perez-Gimenez, Facundo

2015-05-01

The features and theoretical background of a new and free computational program for chemometric analysis denominated IMMAN (acronym for Information theory-based CheMoMetrics ANalysis) are presented. This is multi-platform software developed in the Java programming language, designed with a remarkably user-friendly graphical interface for the computation of a collection of information-theoretic functions adapted for rank-based unsupervised and supervised feature selection tasks. A total of 20 feature selection parameters are presented, with the unsupervised and supervised frameworks represented by 10 approaches in each case. Several information-theoretic parameters traditionally used as molecular descriptors (MDs) are adapted for use as unsupervised rank-based feature selection methods. On the other hand, a generalization scheme for the previously defined differential Shannon's entropy is discussed, as well as the introduction of Jeffreys information measure for supervised feature selection. Moreover, well-known information-theoretic feature selection parameters, such as information gain, gain ratio, and symmetrical uncertainty are incorporated to the IMMAN software ( http://mobiosd-hub.com/imman-soft/ ), following an equal-interval discretization approach. IMMAN offers data pre-processing functionalities, such as missing values processing, dataset partitioning, and browsing. Moreover, single parameter or ensemble (multi-criteria) ranking options are provided. Consequently, this software is suitable for tasks like dimensionality reduction, feature ranking, as well as comparative diversity analysis of data matrices. Simple examples of applications performed with this program are presented. A comparative study between IMMAN and WEKA feature selection tools using the Arcene dataset was performed, demonstrating similar behavior. In addition, it is revealed that the use of IMMAN unsupervised feature selection methods improves the performance of both IMMAN and WEKA supervised algorithms. Graphic representation for Shannon's distribution of MD calculating software.

Kernel Regression Estimation of Fiber Orientation Mixtures in Diffusion MRI

PubMed Central

Cabeen, Ryan P.; Bastin, Mark E.; Laidlaw, David H.

2016-01-01

We present and evaluate a method for kernel regression estimation of fiber orientations and associated volume fractions for diffusion MR tractography and population-based atlas construction in clinical imaging studies of brain white matter. This is a model-based image processing technique in which representative fiber models are estimated from collections of component fiber models in model-valued image data. This extends prior work in nonparametric image processing and multi-compartment processing to provide computational tools for image interpolation, smoothing, and fusion with fiber orientation mixtures. In contrast to related work on multi-compartment processing, this approach is based on directional measures of divergence and includes data-adaptive extensions for model selection and bilateral filtering. This is useful for reconstructing complex anatomical features in clinical datasets analyzed with the ball-and-sticks model, and our framework’s data-adaptive extensions are potentially useful for general multi-compartment image processing. We experimentally evaluate our approach with both synthetic data from computational phantoms and in vivo clinical data from human subjects. With synthetic data experiments, we evaluate performance based on errors in fiber orientation, volume fraction, compartment count, and tractography-based connectivity. With in vivo data experiments, we first show improved scan-rescan reproducibility and reliability of quantitative fiber bundle metrics, including mean length, volume, streamline count, and mean volume fraction. We then demonstrate the creation of a multi-fiber tractography atlas from a population of 80 human subjects. In comparison to single tensor atlasing, our multi-fiber atlas shows more complete features of known fiber bundles and includes reconstructions of the lateral projections of the corpus callosum and complex fronto-parietal connections of the superior longitudinal fasciculus I, II, and III. PMID:26691524

Feature selection for elderly faller classification based on wearable sensors.

PubMed

Howcroft, Jennifer; Kofman, Jonathan; Lemaire, Edward D

2017-05-30

Wearable sensors can be used to derive numerous gait pattern features for elderly fall risk and faller classification; however, an appropriate feature set is required to avoid high computational costs and the inclusion of irrelevant features. The objectives of this study were to identify and evaluate smaller feature sets for faller classification from large feature sets derived from wearable accelerometer and pressure-sensing insole gait data. A convenience sample of 100 older adults (75.5 ± 6.7 years; 76 non-fallers, 24 fallers based on 6 month retrospective fall occurrence) walked 7.62 m while wearing pressure-sensing insoles and tri-axial accelerometers at the head, pelvis, left and right shanks. Feature selection was performed using correlation-based feature selection (CFS), fast correlation based filter (FCBF), and Relief-F algorithms. Faller classification was performed using multi-layer perceptron neural network, naïve Bayesian, and support vector machine classifiers, with 75:25 single stratified holdout and repeated random sampling. The best performing model was a support vector machine with 78% accuracy, 26% sensitivity, 95% specificity, 0.36 F1 score, and 0.31 MCC and one posterior pelvis accelerometer input feature (left acceleration standard deviation). The second best model achieved better sensitivity (44%) and used a support vector machine with 74% accuracy, 83% specificity, 0.44 F1 score, and 0.29 MCC. This model had ten input features: maximum, mean and standard deviation posterior acceleration; maximum, mean and standard deviation anterior acceleration; mean superior acceleration; and three impulse features. The best multi-sensor model sensitivity (56%) was achieved using posterior pelvis and both shank accelerometers and a naïve Bayesian classifier. The best single-sensor model sensitivity (41%) was achieved using the posterior pelvis accelerometer and a naïve Bayesian classifier. Feature selection provided models with smaller feature sets and improved faller classification compared to faller classification without feature selection. CFS and FCBF provided the best feature subset (one posterior pelvis accelerometer feature) for faller classification. However, better sensitivity was achieved by the second best model based on a Relief-F feature subset with three pressure-sensing insole features and seven head accelerometer features. Feature selection should be considered as an important step in faller classification using wearable sensors.

Prediction of cognitive and motor development in preterm children using exhaustive feature selection and cross-validation of near-term white matter microstructure.

PubMed

Schadl, Kornél; Vassar, Rachel; Cahill-Rowley, Katelyn; Yeom, Kristin W; Stevenson, David K; Rose, Jessica

2018-01-01

Advanced neuroimaging and computational methods offer opportunities for more accurate prognosis. We hypothesized that near-term regional white matter (WM) microstructure, assessed on diffusion tensor imaging (DTI), using exhaustive feature selection with cross-validation would predict neurodevelopment in preterm children. Near-term MRI and DTI obtained at 36.6 ± 1.8 weeks postmenstrual age in 66 very-low-birth-weight preterm neonates were assessed. 60/66 had follow-up neurodevelopmental evaluation with Bayley Scales of Infant-Toddler Development, 3rd-edition (BSID-III) at 18-22 months. Linear models with exhaustive feature selection and leave-one-out cross-validation computed based on DTI identified sets of three brain regions most predictive of cognitive and motor function; logistic regression models were computed to classify high-risk infants scoring one standard deviation below mean. Cognitive impairment was predicted (100% sensitivity, 100% specificity; AUC = 1) by near-term right middle-temporal gyrus MD, right cingulate-cingulum MD, left caudate MD. Motor impairment was predicted (90% sensitivity, 86% specificity; AUC = 0.912) by left precuneus FA, right superior occipital gyrus MD, right hippocampus FA. Cognitive score variance was explained (29.6%, cross-validated Rˆ2 = 0.296) by left posterior-limb-of-internal-capsule MD, Genu RD, right fusiform gyrus AD. Motor score variance was explained (31.7%, cross-validated Rˆ2 = 0.317) by left posterior-limb-of-internal-capsule MD, right parahippocampal gyrus AD, right middle-temporal gyrus AD. Search in large DTI feature space more accurately identified neonatal neuroimaging correlates of neurodevelopment.

Enriching 3D optical surface scans with prior knowledge: tissue thickness computation by exploiting local neighborhoods.

PubMed

Wissel, Tobias; Stüber, Patrick; Wagner, Benjamin; Bruder, Ralf; Schweikard, Achim; Ernst, Floris

2016-04-01

Patient immobilization and X-ray-based imaging provide neither a convenient nor a very accurate way to ensure low repositioning errors or to compensate for motion in cranial radiotherapy. We therefore propose an optical tracking device that exploits subcutaneous structures as landmarks in addition to merely spatial registration. To develop such head tracking algorithms, precise and robust computation of these structures is necessary. Here, we show that the tissue thickness can be predicted with high accuracy and moreover exploit local neighborhood information within the laser spot grid on the forehead to further increase this estimation accuracy. We use statistical learning with Support Vector Regression and Gaussian Processes to learn a relationship between optical backscatter features and an MR tissue thickness ground truth. We compare different kernel functions for the data of five different subjects. The incident angle of the laser on the forehead as well as local neighborhoods is incorporated into the feature space. The latter represent the backscatter features from four neighboring laser spots. We confirm that the incident angle has a positive effect on the estimation error of the tissue thickness. The root-mean-square error falls even below 0.15 mm when adding the complete neighborhood information. This prior knowledge also leads to a smoothing effect on the reconstructed skin patch. Learning between different head poses yields similar results. The partial overlap of the point clouds makes the trade-off between novel information and increased feature space dimension obvious and hence feature selection by e.g., sequential forward selection necessary.

Cancer survival classification using integrated data sets and intermediate information.

PubMed

Kim, Shinuk; Park, Taesung; Kon, Mark

2014-09-01

Although numerous studies related to cancer survival have been published, increasing the prediction accuracy of survival classes still remains a challenge. Integration of different data sets, such as microRNA (miRNA) and mRNA, might increase the accuracy of survival class prediction. Therefore, we suggested a machine learning (ML) approach to integrate different data sets, and developed a novel method based on feature selection with Cox proportional hazard regression model (FSCOX) to improve the prediction of cancer survival time. FSCOX provides us with intermediate survival information, which is usually discarded when separating survival into 2 groups (short- and long-term), and allows us to perform survival analysis. We used an ML-based protocol for feature selection, integrating information from miRNA and mRNA expression profiles at the feature level. To predict survival phenotypes, we used the following classifiers, first, existing ML methods, support vector machine (SVM) and random forest (RF), second, a new median-based classifier using FSCOX (FSCOX_median), and third, an SVM classifier using FSCOX (FSCOX_SVM). We compared these methods using 3 types of cancer tissue data sets: (i) miRNA expression, (ii) mRNA expression, and (iii) combined miRNA and mRNA expression. The latter data set included features selected either from the combined miRNA/mRNA profile or independently from miRNAs and mRNAs profiles (IFS). In the ovarian data set, the accuracy of survival classification using the combined miRNA/mRNA profiles with IFS was 75% using RF, 86.36% using SVM, 84.09% using FSCOX_median, and 88.64% using FSCOX_SVM with a balanced 22 short-term and 22 long-term survivor data set. These accuracies are higher than those using miRNA alone (70.45%, RF; 75%, SVM; 75%, FSCOX_median; and 75%, FSCOX_SVM) or mRNA alone (65.91%, RF; 63.64%, SVM; 72.73%, FSCOX_median; and 70.45%, FSCOX_SVM). Similarly in the glioblastoma multiforme data, the accuracy of miRNA/mRNA using IFS was 75.51% (RF), 87.76% (SVM) 85.71% (FSCOX_median), 85.71% (FSCOX_SVM). These results are higher than the results of using miRNA expression and mRNA expression alone. In addition we predict 16 hsa-miR-23b and hsa-miR-27b target genes in ovarian cancer data sets, obtained by SVM-based feature selection through integration of sequence information and gene expression profiles. Among the approaches used, the integrated miRNA and mRNA data set yielded better results than the individual data sets. The best performance was achieved using the FSCOX_SVM method with independent feature selection, which uses intermediate survival information between short-term and long-term survival time and the combination of the 2 different data sets. The results obtained using the combined data set suggest that there are some strong interactions between miRNA and mRNA features that are not detectable in the individual analyses. Copyright © 2014 Elsevier B.V. All rights reserved.

A simple approach to quantitative analysis using three-dimensional spectra based on selected Zernike moments.

PubMed

Zhai, Hong Lin; Zhai, Yue Yuan; Li, Pei Zhen; Tian, Yue Li

2013-01-21

A very simple approach to quantitative analysis is proposed based on the technology of digital image processing using three-dimensional (3D) spectra obtained by high-performance liquid chromatography coupled with a diode array detector (HPLC-DAD). As the region-based shape features of a grayscale image, Zernike moments with inherently invariance property were employed to establish the linear quantitative models. This approach was applied to the quantitative analysis of three compounds in mixed samples using 3D HPLC-DAD spectra, and three linear models were obtained, respectively. The correlation coefficients (R(2)) for training and test sets were more than 0.999, and the statistical parameters and strict validation supported the reliability of established models. The analytical results suggest that the Zernike moment selected by stepwise regression can be used in the quantitative analysis of target compounds. Our study provides a new idea for quantitative analysis using 3D spectra, which can be extended to the analysis of other 3D spectra obtained by different methods or instruments.

A hybrid feature selection method using multiclass SVM for diagnosis of erythemato-squamous disease

NASA Astrophysics Data System (ADS)

Maryam, Setiawan, Noor Akhmad; Wahyunggoro, Oyas

2017-08-01

The diagnosis of erythemato-squamous disease is a complex problem and difficult to detect in dermatology. Besides that, it is a major cause of skin cancer. Data mining implementation in the medical field helps expert to diagnose precisely, accurately, and inexpensively. In this research, we use data mining technique to developed a diagnosis model based on multiclass SVM with a novel hybrid feature selection method to diagnose erythemato-squamous disease. Our hybrid feature selection method, named ChiGA (Chi Square and Genetic Algorithm), uses the advantages from filter and wrapper methods to select the optimal feature subset from original feature. Chi square used as filter method to remove redundant features and GA as wrapper method to select the ideal feature subset with SVM used as classifier. Experiment performed with 10 fold cross validation on erythemato-squamous diseases dataset taken from University of California Irvine (UCI) machine learning database. The experimental result shows that the proposed model based multiclass SVM with Chi Square and GA can give an optimum feature subset. There are 18 optimum features with 99.18% accuracy.

Constant size descriptors for accurate machine learning models of molecular properties

NASA Astrophysics Data System (ADS)

Collins, Christopher R.; Gordon, Geoffrey J.; von Lilienfeld, O. Anatole; Yaron, David J.

2018-06-01

Two different classes of molecular representations for use in machine learning of thermodynamic and electronic properties are studied. The representations are evaluated by monitoring the performance of linear and kernel ridge regression models on well-studied data sets of small organic molecules. One class of representations studied here counts the occurrence of bonding patterns in the molecule. These require only the connectivity of atoms in the molecule as may be obtained from a line diagram or a SMILES string. The second class utilizes the three-dimensional structure of the molecule. These include the Coulomb matrix and Bag of Bonds, which list the inter-atomic distances present in the molecule, and Encoded Bonds, which encode such lists into a feature vector whose length is independent of molecular size. Encoded Bonds' features introduced here have the advantage of leading to models that may be trained on smaller molecules and then used successfully on larger molecules. A wide range of feature sets are constructed by selecting, at each rank, either a graph or geometry-based feature. Here, rank refers to the number of atoms involved in the feature, e.g., atom counts are rank 1, while Encoded Bonds are rank 2. For atomization energies in the QM7 data set, the best graph-based feature set gives a mean absolute error of 3.4 kcal/mol. Inclusion of 3D geometry substantially enhances the performance, with Encoded Bonds giving 2.4 kcal/mol, when used alone, and 1.19 kcal/mol, when combined with graph features.

A Feature-Free 30-Disease Pathological Brain Detection System by Linear Regression Classifier.

PubMed

Chen, Yi; Shao, Ying; Yan, Jie; Yuan, Ti-Fei; Qu, Yanwen; Lee, Elizabeth; Wang, Shuihua

2017-01-01

Alzheimer's disease patients are increasing rapidly every year. Scholars tend to use computer vision methods to develop automatic diagnosis system. (Background) In 2015, Gorji et al. proposed a novel method using pseudo Zernike moment. They tested four classifiers: learning vector quantization neural network, pattern recognition neural network trained by Levenberg-Marquardt, by resilient backpropagation, and by scaled conjugate gradient. This study presents an improved method by introducing a relatively new classifier-linear regression classification. Our method selects one axial slice from 3D brain image, and employed pseudo Zernike moment with maximum order of 15 to extract 256 features from each image. Finally, linear regression classification was harnessed as the classifier. The proposed approach obtains an accuracy of 97.51%, a sensitivity of 96.71%, and a specificity of 97.73%. Our method performs better than Gorji's approach and five other state-of-the-art approaches. Therefore, it can be used to detect Alzheimer's disease. Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.org.

Feature selection for wearable smartphone-based human activity recognition with able bodied, elderly, and stroke patients.

PubMed

Capela, Nicole A; Lemaire, Edward D; Baddour, Natalie

2015-01-01

Human activity recognition (HAR), using wearable sensors, is a growing area with the potential to provide valuable information on patient mobility to rehabilitation specialists. Smartphones with accelerometer and gyroscope sensors are a convenient, minimally invasive, and low cost approach for mobility monitoring. HAR systems typically pre-process raw signals, segment the signals, and then extract features to be used in a classifier. Feature selection is a crucial step in the process to reduce potentially large data dimensionality and provide viable parameters to enable activity classification. Most HAR systems are customized to an individual research group, including a unique data set, classes, algorithms, and signal features. These data sets are obtained predominantly from able-bodied participants. In this paper, smartphone accelerometer and gyroscope sensor data were collected from populations that can benefit from human activity recognition: able-bodied, elderly, and stroke patients. Data from a consecutive sequence of 41 mobility tasks (18 different tasks) were collected for a total of 44 participants. Seventy-six signal features were calculated and subsets of these features were selected using three filter-based, classifier-independent, feature selection methods (Relief-F, Correlation-based Feature Selection, Fast Correlation Based Filter). The feature subsets were then evaluated using three generic classifiers (Naïve Bayes, Support Vector Machine, j48 Decision Tree). Common features were identified for all three populations, although the stroke population subset had some differences from both able-bodied and elderly sets. Evaluation with the three classifiers showed that the feature subsets produced similar or better accuracies than classification with the entire feature set. Therefore, since these feature subsets are classifier-independent, they should be useful for developing and improving HAR systems across and within populations.

Feature Selection for Wearable Smartphone-Based Human Activity Recognition with Able bodied, Elderly, and Stroke Patients

PubMed Central

2015-01-01

Human activity recognition (HAR), using wearable sensors, is a growing area with the potential to provide valuable information on patient mobility to rehabilitation specialists. Smartphones with accelerometer and gyroscope sensors are a convenient, minimally invasive, and low cost approach for mobility monitoring. HAR systems typically pre-process raw signals, segment the signals, and then extract features to be used in a classifier. Feature selection is a crucial step in the process to reduce potentially large data dimensionality and provide viable parameters to enable activity classification. Most HAR systems are customized to an individual research group, including a unique data set, classes, algorithms, and signal features. These data sets are obtained predominantly from able-bodied participants. In this paper, smartphone accelerometer and gyroscope sensor data were collected from populations that can benefit from human activity recognition: able-bodied, elderly, and stroke patients. Data from a consecutive sequence of 41 mobility tasks (18 different tasks) were collected for a total of 44 participants. Seventy-six signal features were calculated and subsets of these features were selected using three filter-based, classifier-independent, feature selection methods (Relief-F, Correlation-based Feature Selection, Fast Correlation Based Filter). The feature subsets were then evaluated using three generic classifiers (Naïve Bayes, Support Vector Machine, j48 Decision Tree). Common features were identified for all three populations, although the stroke population subset had some differences from both able-bodied and elderly sets. Evaluation with the three classifiers showed that the feature subsets produced similar or better accuracies than classification with the entire feature set. Therefore, since these feature subsets are classifier-independent, they should be useful for developing and improving HAR systems across and within populations. PMID:25885272

Estimating Influenza Outbreaks Using Both Search Engine Query Data and Social Media Data in South Korea.

PubMed

Woo, Hyekyung; Cho, Youngtae; Shim, Eunyoung; Lee, Jong-Koo; Lee, Chang-Gun; Kim, Seong Hwan

2016-07-04

As suggested as early as in 2006, logs of queries submitted to search engines seeking information could be a source for detection of emerging influenza epidemics if changes in the volume of search queries are monitored (infodemiology). However, selecting queries that are most likely to be associated with influenza epidemics is a particular challenge when it comes to generating better predictions. In this study, we describe a methodological extension for detecting influenza outbreaks using search query data; we provide a new approach for query selection through the exploration of contextual information gleaned from social media data. Additionally, we evaluate whether it is possible to use these queries for monitoring and predicting influenza epidemics in South Korea. Our study was based on freely available weekly influenza incidence data and query data originating from the search engine on the Korean website Daum between April 3, 2011 and April 5, 2014. To select queries related to influenza epidemics, several approaches were applied: (1) exploring influenza-related words in social media data, (2) identifying the chief concerns related to influenza, and (3) using Web query recommendations. Optimal feature selection by least absolute shrinkage and selection operator (Lasso) and support vector machine for regression (SVR) were used to construct a model predicting influenza epidemics. In total, 146 queries related to influenza were generated through our initial query selection approach. A considerable proportion of optimal features for final models were derived from queries with reference to the social media data. The SVR model performed well: the prediction values were highly correlated with the recent observed influenza-like illness (r=.956; P<.001) and virological incidence rate (r=.963; P<.001). These results demonstrate the feasibility of using search queries to enhance influenza surveillance in South Korea. In addition, an approach for query selection using social media data seems ideal for supporting influenza surveillance based on search query data.

Estimating Influenza Outbreaks Using Both Search Engine Query Data and Social Media Data in South Korea

PubMed Central

Woo, Hyekyung; Shim, Eunyoung; Lee, Jong-Koo; Lee, Chang-Gun; Kim, Seong Hwan

2016-01-01

Background As suggested as early as in 2006, logs of queries submitted to search engines seeking information could be a source for detection of emerging influenza epidemics if changes in the volume of search queries are monitored (infodemiology). However, selecting queries that are most likely to be associated with influenza epidemics is a particular challenge when it comes to generating better predictions. Objective In this study, we describe a methodological extension for detecting influenza outbreaks using search query data; we provide a new approach for query selection through the exploration of contextual information gleaned from social media data. Additionally, we evaluate whether it is possible to use these queries for monitoring and predicting influenza epidemics in South Korea. Methods Our study was based on freely available weekly influenza incidence data and query data originating from the search engine on the Korean website Daum between April 3, 2011 and April 5, 2014. To select queries related to influenza epidemics, several approaches were applied: (1) exploring influenza-related words in social media data, (2) identifying the chief concerns related to influenza, and (3) using Web query recommendations. Optimal feature selection by least absolute shrinkage and selection operator (Lasso) and support vector machine for regression (SVR) were used to construct a model predicting influenza epidemics. Results In total, 146 queries related to influenza were generated through our initial query selection approach. A considerable proportion of optimal features for final models were derived from queries with reference to the social media data. The SVR model performed well: the prediction values were highly correlated with the recent observed influenza-like illness (r=.956; P<.001) and virological incidence rate (r=.963; P<.001). Conclusions These results demonstrate the feasibility of using search queries to enhance influenza surveillance in South Korea. In addition, an approach for query selection using social media data seems ideal for supporting influenza surveillance based on search query data. PMID:27377323

Feature Selection Method Based on Neighborhood Relationships: Applications in EEG Signal Identification and Chinese Character Recognition

PubMed Central

Zhao, Yu-Xiang; Chou, Chien-Hsing

2016-01-01

In this study, a new feature selection algorithm, the neighborhood-relationship feature selection (NRFS) algorithm, is proposed for identifying rat electroencephalogram signals and recognizing Chinese characters. In these two applications, dependent relationships exist among the feature vectors and their neighboring feature vectors. Therefore, the proposed NRFS algorithm was designed for solving this problem. By applying the NRFS algorithm, unselected feature vectors have a high priority of being added into the feature subset if the neighboring feature vectors have been selected. In addition, selected feature vectors have a high priority of being eliminated if the neighboring feature vectors are not selected. In the experiments conducted in this study, the NRFS algorithm was compared with two feature algorithms. The experimental results indicated that the NRFS algorithm can extract the crucial frequency bands for identifying rat vigilance states and identifying crucial character regions for recognizing Chinese characters. PMID:27314346

Compact cancer biomarkers discovery using a swarm intelligence feature selection algorithm.

PubMed

Martinez, Emmanuel; Alvarez, Mario Moises; Trevino, Victor

2010-08-01

Biomarker discovery is a typical application from functional genomics. Due to the large number of genes studied simultaneously in microarray data, feature selection is a key step. Swarm intelligence has emerged as a solution for the feature selection problem. However, swarm intelligence settings for feature selection fail to select small features subsets. We have proposed a swarm intelligence feature selection algorithm based on the initialization and update of only a subset of particles in the swarm. In this study, we tested our algorithm in 11 microarray datasets for brain, leukemia, lung, prostate, and others. We show that the proposed swarm intelligence algorithm successfully increase the classification accuracy and decrease the number of selected features compared to other swarm intelligence methods. Copyright © 2010 Elsevier Ltd. All rights reserved.

Computational intelligence models to predict porosity of tablets using minimum features

PubMed Central

Khalid, Mohammad Hassan; Kazemi, Pezhman; Perez-Gandarillas, Lucia; Michrafy, Abderrahim; Szlęk, Jakub; Jachowicz, Renata; Mendyk, Aleksander

2017-01-01

The effects of different formulations and manufacturing process conditions on the physical properties of a solid dosage form are of importance to the pharmaceutical industry. It is vital to have in-depth understanding of the material properties and governing parameters of its processes in response to different formulations. Understanding the mentioned aspects will allow tighter control of the process, leading to implementation of quality-by-design (QbD) practices. Computational intelligence (CI) offers an opportunity to create empirical models that can be used to describe the system and predict future outcomes in silico. CI models can help explore the behavior of input parameters, unlocking deeper understanding of the system. This research endeavor presents CI models to predict the porosity of tablets created by roll-compacted binary mixtures, which were milled and compacted under systematically varying conditions. CI models were created using tree-based methods, artificial neural networks (ANNs), and symbolic regression trained on an experimental data set and screened using root-mean-square error (RMSE) scores. The experimental data were composed of proportion of microcrystalline cellulose (MCC) (in percentage), granule size fraction (in micrometers), and die compaction force (in kilonewtons) as inputs and porosity as an output. The resulting models show impressive generalization ability, with ANNs (normalized root-mean-square error [NRMSE] =1%) and symbolic regression (NRMSE =4%) as the best-performing methods, also exhibiting reliable predictive behavior when presented with a challenging external validation data set (best achieved symbolic regression: NRMSE =3%). Symbolic regression demonstrates the transition from the black box modeling paradigm to more transparent predictive models. Predictive performance and feature selection behavior of CI models hints at the most important variables within this factor space. PMID:28138223

Computational intelligence models to predict porosity of tablets using minimum features.

PubMed

Khalid, Mohammad Hassan; Kazemi, Pezhman; Perez-Gandarillas, Lucia; Michrafy, Abderrahim; Szlęk, Jakub; Jachowicz, Renata; Mendyk, Aleksander

2017-01-01

The effects of different formulations and manufacturing process conditions on the physical properties of a solid dosage form are of importance to the pharmaceutical industry. It is vital to have in-depth understanding of the material properties and governing parameters of its processes in response to different formulations. Understanding the mentioned aspects will allow tighter control of the process, leading to implementation of quality-by-design (QbD) practices. Computational intelligence (CI) offers an opportunity to create empirical models that can be used to describe the system and predict future outcomes in silico. CI models can help explore the behavior of input parameters, unlocking deeper understanding of the system. This research endeavor presents CI models to predict the porosity of tablets created by roll-compacted binary mixtures, which were milled and compacted under systematically varying conditions. CI models were created using tree-based methods, artificial neural networks (ANNs), and symbolic regression trained on an experimental data set and screened using root-mean-square error (RMSE) scores. The experimental data were composed of proportion of microcrystalline cellulose (MCC) (in percentage), granule size fraction (in micrometers), and die compaction force (in kilonewtons) as inputs and porosity as an output. The resulting models show impressive generalization ability, with ANNs (normalized root-mean-square error [NRMSE] =1%) and symbolic regression (NRMSE =4%) as the best-performing methods, also exhibiting reliable predictive behavior when presented with a challenging external validation data set (best achieved symbolic regression: NRMSE =3%). Symbolic regression demonstrates the transition from the black box modeling paradigm to more transparent predictive models. Predictive performance and feature selection behavior of CI models hints at the most important variables within this factor space.

Classification of independent components of EEG into multiple artifact classes.

PubMed

Frølich, Laura; Andersen, Tobias S; Mørup, Morten

2015-01-01

In this study, we aim to automatically identify multiple artifact types in EEG. We used multinomial regression to classify independent components of EEG data, selecting from 65 spatial, spectral, and temporal features of independent components using forward selection. The classifier identified neural and five nonneural types of components. Between subjects within studies, high classification performances were obtained. Between studies, however, classification was more difficult. For neural versus nonneural classifications, performance was on par with previous results obtained by others. We found that automatic separation of multiple artifact classes is possible with a small feature set. Our method can reduce manual workload and allow for the selective removal of artifact classes. Identifying artifacts during EEG recording may be used to instruct subjects to refrain from activity causing them. Copyright © 2014 Society for Psychophysiological Research.

VARSEDIG: an algorithm for morphometric characters selection and statistical validation in morphological taxonomy.

PubMed

Guisande, Cástor; Vari, Richard P; Heine, Jürgen; García-Roselló, Emilio; González-Dacosta, Jacinto; Perez-Schofield, Baltasar J García; González-Vilas, Luis; Pelayo-Villamil, Patricia

2016-09-12

We present and discuss VARSEDIG, an algorithm which identifies the morphometric features that significantly discriminate two taxa and validates the morphological distinctness between them via a Monte-Carlo test. VARSEDIG is freely available as a function of the RWizard application PlotsR (http://www.ipez.es/RWizard) and as R package on CRAN. The variables selected by VARSEDIG with the overlap method were very similar to those selected by logistic regression and discriminant analysis, but overcomes some shortcomings of these methods. VARSEDIG is, therefore, a good alternative by comparison to current classical classification methods for identifying morphometric features that significantly discriminate a taxon and for validating its morphological distinctness from other taxa. As a demonstration of the potential of VARSEDIG for this purpose, we analyze morphological discrimination among some species of the Neotropical freshwater family Characidae.

Computational Prediction of Protein Epsilon Lysine Acetylation Sites Based on a Feature Selection Method.

PubMed

Gao, JianZhao; Tao, Xue-Wen; Zhao, Jia; Feng, Yuan-Ming; Cai, Yu-Dong; Zhang, Ning

2017-01-01

Lysine acetylation, as one type of post-translational modifications (PTM), plays key roles in cellular regulations and can be involved in a variety of human diseases. However, it is often high-cost and time-consuming to use traditional experimental approaches to identify the lysine acetylation sites. Therefore, effective computational methods should be developed to predict the acetylation sites. In this study, we developed a position-specific method for epsilon lysine acetylation site prediction. Sequences of acetylated proteins were retrieved from the UniProt database. Various kinds of features such as position specific scoring matrix (PSSM), amino acid factors (AAF), and disorders were incorporated. A feature selection method based on mRMR (Maximum Relevance Minimum Redundancy) and IFS (Incremental Feature Selection) was employed. Finally, 319 optimal features were selected from total 541 features. Using the 319 optimal features to encode peptides, a predictor was constructed based on dagging. As a result, an accuracy of 69.56% with MCC of 0.2792 was achieved. We analyzed the optimal features, which suggested some important factors determining the lysine acetylation sites. We developed a position-specific method for epsilon lysine acetylation site prediction. A set of optimal features was selected. Analysis of the optimal features provided insights into the mechanism of lysine acetylation sites, providing guidance of experimental validation. Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.org.

Minimizing the semantic gap in biomedical content-based image retrieval

NASA Astrophysics Data System (ADS)

Guan, Haiying; Antani, Sameer; Long, L. Rodney; Thoma, George R.

2010-03-01

A major challenge in biomedical Content-Based Image Retrieval (CBIR) is to achieve meaningful mappings that minimize the semantic gap between the high-level biomedical semantic concepts and the low-level visual features in images. This paper presents a comprehensive learning-based scheme toward meeting this challenge and improving retrieval quality. The article presents two algorithms: a learning-based feature selection and fusion algorithm and the Ranking Support Vector Machine (Ranking SVM) algorithm. The feature selection algorithm aims to select 'good' features and fuse them using different similarity measurements to provide a better representation of the high-level concepts with the low-level image features. Ranking SVM is applied to learn the retrieval rank function and associate the selected low-level features with query concepts, given the ground-truth ranking of the training samples. The proposed scheme addresses four major issues in CBIR to improve the retrieval accuracy: image feature extraction, selection and fusion, similarity measurements, the association of the low-level features with high-level concepts, and the generation of the rank function to support high-level semantic image retrieval. It models the relationship between semantic concepts and image features, and enables retrieval at the semantic level. We apply it to the problem of vertebra shape retrieval from a digitized spine x-ray image set collected by the second National Health and Nutrition Examination Survey (NHANES II). The experimental results show an improvement of up to 41.92% in the mean average precision (MAP) over conventional image similarity computation methods.

Classification of Medical Datasets Using SVMs with Hybrid Evolutionary Algorithms Based on Endocrine-Based Particle Swarm Optimization and Artificial Bee Colony Algorithms.

PubMed

Lin, Kuan-Cheng; Hsieh, Yi-Hsiu

2015-10-01

The classification and analysis of data is an important issue in today's research. Selecting a suitable set of features makes it possible to classify an enormous quantity of data quickly and efficiently. Feature selection is generally viewed as a problem of feature subset selection, such as combination optimization problems. Evolutionary algorithms using random search methods have proven highly effective in obtaining solutions to problems of optimization in a diversity of applications. In this study, we developed a hybrid evolutionary algorithm based on endocrine-based particle swarm optimization (EPSO) and artificial bee colony (ABC) algorithms in conjunction with a support vector machine (SVM) for the selection of optimal feature subsets for the classification of datasets. The results of experiments using specific UCI medical datasets demonstrate that the accuracy of the proposed hybrid evolutionary algorithm is superior to that of basic PSO, EPSO and ABC algorithms, with regard to classification accuracy using subsets with a reduced number of features.

Feature selection for neural network based defect classification of ceramic components using high frequency ultrasound.

PubMed

Kesharaju, Manasa; Nagarajah, Romesh

2015-09-01

The motivation for this research stems from a need for providing a non-destructive testing method capable of detecting and locating any defects and microstructural variations within armour ceramic components before issuing them to the soldiers who rely on them for their survival. The development of an automated ultrasonic inspection based classification system would make possible the checking of each ceramic component and immediately alert the operator about the presence of defects. Generally, in many classification problems a choice of features or dimensionality reduction is significant and simultaneously very difficult, as a substantial computational effort is required to evaluate possible feature subsets. In this research, a combination of artificial neural networks and genetic algorithms are used to optimize the feature subset used in classification of various defects in reaction-sintered silicon carbide ceramic components. Initially wavelet based feature extraction is implemented from the region of interest. An Artificial Neural Network classifier is employed to evaluate the performance of these features. Genetic Algorithm based feature selection is performed. Principal Component Analysis is a popular technique used for feature selection and is compared with the genetic algorithm based technique in terms of classification accuracy and selection of optimal number of features. The experimental results confirm that features identified by Principal Component Analysis lead to improved performance in terms of classification percentage with 96% than Genetic algorithm with 94%. Copyright © 2015 Elsevier B.V. All rights reserved.

A three-step approach for the derivation and validation of high-performing predictive models using an operational dataset: congestive heart failure readmission case study.

PubMed

AbdelRahman, Samir E; Zhang, Mingyuan; Bray, Bruce E; Kawamoto, Kensaku

2014-05-27

The aim of this study was to propose an analytical approach to develop high-performing predictive models for congestive heart failure (CHF) readmission using an operational dataset with incomplete records and changing data over time. Our analytical approach involves three steps: pre-processing, systematic model development, and risk factor analysis. For pre-processing, variables that were absent in >50% of records were removed. Moreover, the dataset was divided into a validation dataset and derivation datasets which were separated into three temporal subsets based on changes to the data over time. For systematic model development, using the different temporal datasets and the remaining explanatory variables, the models were developed by combining the use of various (i) statistical analyses to explore the relationships between the validation and the derivation datasets; (ii) adjustment methods for handling missing values; (iii) classifiers; (iv) feature selection methods; and (iv) discretization methods. We then selected the best derivation dataset and the models with the highest predictive performance. For risk factor analysis, factors in the highest-performing predictive models were analyzed and ranked using (i) statistical analyses of the best derivation dataset, (ii) feature rankers, and (iii) a newly developed algorithm to categorize risk factors as being strong, regular, or weak. The analysis dataset consisted of 2,787 CHF hospitalizations at University of Utah Health Care from January 2003 to June 2013. In this study, we used the complete-case analysis and mean-based imputation adjustment methods; the wrapper subset feature selection method; and four ranking strategies based on information gain, gain ratio, symmetrical uncertainty, and wrapper subset feature evaluators. The best-performing models resulted from the use of a complete-case analysis derivation dataset combined with the Class-Attribute Contingency Coefficient discretization method and a voting classifier which averaged the results of multi-nominal logistic regression and voting feature intervals classifiers. Of 42 final model risk factors, discharge disposition, discretized age, and indicators of anemia were the most significant. This model achieved a c-statistic of 86.8%. The proposed three-step analytical approach enhanced predictive model performance for CHF readmissions. It could potentially be leveraged to improve predictive model performance in other areas of clinical medicine.

Improved Measurement of Blood Pressure by Extraction of Characteristic Features from the Cuff Oscillometric Waveform

PubMed Central

Lim, Pooi Khoon; Ng, Siew-Cheok; Jassim, Wissam A.; Redmond, Stephen J.; Zilany, Mohammad; Avolio, Alberto; Lim, Einly; Tan, Maw Pin; Lovell, Nigel H.

2015-01-01

We present a novel approach to improve the estimation of systolic (SBP) and diastolic blood pressure (DBP) from oscillometric waveform data using variable characteristic ratios between SBP and DBP with mean arterial pressure (MAP). This was verified in 25 healthy subjects, aged 28 ± 5 years. The multiple linear regression (MLR) and support vector regression (SVR) models were used to examine the relationship between the SBP and the DBP ratio with ten features extracted from the oscillometric waveform envelope (OWE). An automatic algorithm based on relative changes in the cuff pressure and neighbouring oscillometric pulses was proposed to remove outlier points caused by movement artifacts. Substantial reduction in the mean and standard deviation of the blood pressure estimation errors were obtained upon artifact removal. Using the sequential forward floating selection (SFFS) approach, we were able to achieve a significant reduction in the mean and standard deviation of differences between the estimated SBP values and the reference scoring (MLR: mean ± SD = −0.3 ± 5.8 mmHg; SVR and −0.6 ± 5.4 mmHg) with only two features, i.e., Ratio2 and Area3, as compared to the conventional maximum amplitude algorithm (MAA) method (mean ± SD = −1.6 ± 8.6 mmHg). Comparing the performance of both MLR and SVR models, our results showed that the MLR model was able to achieve comparable performance to that of the SVR model despite its simplicity. PMID:26087370

Feature-based and spatial attentional selection in visual working memory.

PubMed

Heuer, Anna; Schubö, Anna

2016-05-01

The contents of visual working memory (VWM) can be modulated by spatial cues presented during the maintenance interval ("retrocues"). Here, we examined whether attentional selection of representations in VWM can also be based on features. In addition, we investigated whether the mechanisms of feature-based and spatial attention in VWM differ with respect to parallel access to noncontiguous locations. In two experiments, we tested the efficacy of valid retrocues relying on different kinds of information. Specifically, participants were presented with a typical spatial retrocue pointing to two locations, a symbolic spatial retrocue (numbers mapping onto two locations), and two feature-based retrocues: a color retrocue (a blob of the same color as two of the items) and a shape retrocue (an outline of the shape of two of the items). The two cued items were presented at either contiguous or noncontiguous locations. Overall retrocueing benefits, as compared to a neutral condition, were observed for all retrocue types. Whereas feature-based retrocues yielded benefits for cued items presented at both contiguous and noncontiguous locations, spatial retrocues were only effective when the cued items had been presented at contiguous locations. These findings demonstrate that attentional selection and updating in VWM can operate on different kinds of information, allowing for a flexible and efficient use of this limited system. The observation that the representations of items presented at noncontiguous locations could only be reliably selected with feature-based retrocues suggests that feature-based and spatial attentional selection in VWM rely on different mechanisms, as has been shown for attentional orienting in the external world.

A New Approach for Mobile Advertising Click-Through Rate Estimation Based on Deep Belief Nets.

PubMed

Chen, Jie-Hao; Zhao, Zi-Qian; Shi, Ji-Yun; Zhao, Chong

2017-01-01

In recent years, with the rapid development of mobile Internet and its business applications, mobile advertising Click-Through Rate (CTR) estimation has become a hot research direction in the field of computational advertising, which is used to achieve accurate advertisement delivery for the best benefits in the three-side game between media, advertisers, and audiences. Current research on the estimation of CTR mainly uses the methods and models of machine learning, such as linear model or recommendation algorithms. However, most of these methods are insufficient to extract the data features and cannot reflect the nonlinear relationship between different features. In order to solve these problems, we propose a new model based on Deep Belief Nets to predict the CTR of mobile advertising, which combines together the powerful data representation and feature extraction capability of Deep Belief Nets, with the advantage of simplicity of traditional Logistic Regression models. Based on the training dataset with the information of over 40 million mobile advertisements during a period of 10 days, our experiments show that our new model has better estimation accuracy than the classic Logistic Regression (LR) model by 5.57% and Support Vector Regression (SVR) model by 5.80%.

A New Approach for Mobile Advertising Click-Through Rate Estimation Based on Deep Belief Nets

PubMed Central

Zhao, Zi-Qian; Shi, Ji-Yun; Zhao, Chong

2017-01-01

In recent years, with the rapid development of mobile Internet and its business applications, mobile advertising Click-Through Rate (CTR) estimation has become a hot research direction in the field of computational advertising, which is used to achieve accurate advertisement delivery for the best benefits in the three-side game between media, advertisers, and audiences. Current research on the estimation of CTR mainly uses the methods and models of machine learning, such as linear model or recommendation algorithms. However, most of these methods are insufficient to extract the data features and cannot reflect the nonlinear relationship between different features. In order to solve these problems, we propose a new model based on Deep Belief Nets to predict the CTR of mobile advertising, which combines together the powerful data representation and feature extraction capability of Deep Belief Nets, with the advantage of simplicity of traditional Logistic Regression models. Based on the training dataset with the information of over 40 million mobile advertisements during a period of 10 days, our experiments show that our new model has better estimation accuracy than the classic Logistic Regression (LR) model by 5.57% and Support Vector Regression (SVR) model by 5.80%. PMID:29209363

Prediction of Response to Neoadjuvant Chemotherapy and Radiation Therapy with Baseline and Restaging 18F-FDG PET Imaging Biomarkers in Patients with Esophageal Cancer.

PubMed

Beukinga, Roelof J; Hulshoff, Jan Binne; Mul, Véronique E M; Noordzij, Walter; Kats-Ugurlu, Gursah; Slart, Riemer H J A; Plukker, John T M

2018-06-01

Purpose To assess the value of baseline and restaging fluorine 18 ( 18 F) fluorodeoxyglucose (FDG) positron emission tomography (PET) radiomics in predicting pathologic complete response to neoadjuvant chemotherapy and radiation therapy (NCRT) in patients with locally advanced esophageal cancer. Materials and Methods In this retrospective study, 73 patients with histologic analysis-confirmed T1/N1-3/M0 or T2-4a/N0-3/M0 esophageal cancer were treated with NCRT followed by surgery (Chemoradiotherapy for Esophageal Cancer followed by Surgery Study regimen) between October 2014 and August 2017. Clinical variables and radiomic features from baseline and restaging 18 F-FDG PET were selected by univariable logistic regression and least absolute shrinkage and selection operator. The selected variables were used to fit a multivariable logistic regression model, which was internally validated by using bootstrap resampling with 20 000 replicates. The performance of this model was compared with reference prediction models composed of maximum standardized uptake value metrics, clinical variables, and maximum standardized uptake value at baseline NCRT radiomic features. Outcome was defined as complete versus incomplete pathologic response (tumor regression grade 1 vs 2-5 according to the Mandard classification). Results Pathologic response was complete in 16 patients (21.9%) and incomplete in 57 patients (78.1%). A prediction model combining clinical T-stage and restaging NCRT (post-NCRT) joint maximum (quantifying image orderliness) yielded an optimism-corrected area under the receiver operating characteristics curve of 0.81. Post-NCRT joint maximum was replaceable with five other redundant post-NCRT radiomic features that provided equal model performance. All reference prediction models exhibited substantially lower discriminatory accuracy. Conclusion The combination of clinical T-staging and quantitative assessment of post-NCRT 18 F-FDG PET orderliness (joint maximum) provided high discriminatory accuracy in predicting pathologic complete response in patients with esophageal cancer. © RSNA, 2018 Online supplemental material is available for this article.

Automatic migraine classification via feature selection committee and machine learning techniques over imaging and questionnaire data.

PubMed

Garcia-Chimeno, Yolanda; Garcia-Zapirain, Begonya; Gomez-Beldarrain, Marian; Fernandez-Ruanova, Begonya; Garcia-Monco, Juan Carlos

2017-04-13

Feature selection methods are commonly used to identify subsets of relevant features to facilitate the construction of models for classification, yet little is known about how feature selection methods perform in diffusion tensor images (DTIs). In this study, feature selection and machine learning classification methods were tested for the purpose of automating diagnosis of migraines using both DTIs and questionnaire answers related to emotion and cognition - factors that influence of pain perceptions. We select 52 adult subjects for the study divided into three groups: control group (15), subjects with sporadic migraine (19) and subjects with chronic migraine and medication overuse (18). These subjects underwent magnetic resonance with diffusion tensor to see white matter pathway integrity of the regions of interest involved in pain and emotion. The tests also gather data about pathology. The DTI images and test results were then introduced into feature selection algorithms (Gradient Tree Boosting, L1-based, Random Forest and Univariate) to reduce features of the first dataset and classification algorithms (SVM (Support Vector Machine), Boosting (Adaboost) and Naive Bayes) to perform a classification of migraine group. Moreover we implement a committee method to improve the classification accuracy based on feature selection algorithms. When classifying the migraine group, the greatest improvements in accuracy were made using the proposed committee-based feature selection method. Using this approach, the accuracy of classification into three types improved from 67 to 93% when using the Naive Bayes classifier, from 90 to 95% with the support vector machine classifier, 93 to 94% in boosting. The features that were determined to be most useful for classification included are related with the pain, analgesics and left uncinate brain (connected with the pain and emotions). The proposed feature selection committee method improved the performance of migraine diagnosis classifiers compared to individual feature selection methods, producing a robust system that achieved over 90% accuracy in all classifiers. The results suggest that the proposed methods can be used to support specialists in the classification of migraines in patients undergoing magnetic resonance imaging.

Photoacoustic spectroscopy based investigatory approach to discriminate breast cancer from normal: a pilot study

NASA Astrophysics Data System (ADS)

Priya, Mallika; Rao, Bola Sadashiva Satish; Chandra, Subhash; Ray, Satadru; Mathew, Stanley; Datta, Anirbit; Nayak, Subramanya G.; Mahato, Krishna Kishore

2016-02-01

In spite of many efforts for early detection of breast cancer, there is still lack of technology for immediate implementation. In the present study, the potential photoacoustic spectroscopy was evaluated in discriminating breast cancer from normal, involving blood serum samples seeking early detection. Three photoacoustic spectra in time domain were recorded from each of 20 normal and 20 malignant samples at 281nm pulsed laser excitations and a total of 120 spectra were generated. The time domain spectra were then Fast Fourier Transformed into frequency domain and 116.5625 - 206.875 kHz region was selected for further analysis using a combinational approach of wavelet, PCA and logistic regression. Initially, wavelet analysis was performed on the FFT data and seven features (mean, median, area under the curve, variance, standard deviation, skewness and kurtosis) from each were extracted. PCA was then performed on the feature matrix (7x120) for discriminating malignant samples from the normal by plotting a decision boundary using logistic regression analysis. The unsupervised mode of classification used in the present study yielded specificity and sensitivity values of 100% in each respectively with a ROC - AUC value of 1. The results obtained have clearly demonstrated the capability of photoacoustic spectroscopy in discriminating cancer from the normal, suggesting its possible clinical implications.

Effect of feature-selective attention on neuronal responses in macaque area MT

PubMed Central

Chen, X.; Hoffmann, K.-P.; Albright, T. D.

2012-01-01

Attention influences visual processing in striate and extrastriate cortex, which has been extensively studied for spatial-, object-, and feature-based attention. Most studies exploring neural signatures of feature-based attention have trained animals to attend to an object identified by a certain feature and ignore objects/displays identified by a different feature. Little is known about the effects of feature-selective attention, where subjects attend to one stimulus feature domain (e.g., color) of an object while features from different domains (e.g., direction of motion) of the same object are ignored. To study this type of feature-selective attention in area MT in the middle temporal sulcus, we trained macaque monkeys to either attend to and report the direction of motion of a moving sine wave grating (a feature for which MT neurons display strong selectivity) or attend to and report its color (a feature for which MT neurons have very limited selectivity). We hypothesized that neurons would upregulate their firing rate during attend-direction conditions compared with attend-color conditions. We found that feature-selective attention significantly affected 22% of MT neurons. Contrary to our hypothesis, these neurons did not necessarily increase firing rate when animals attended to direction of motion but fell into one of two classes. In one class, attention to color increased the gain of stimulus-induced responses compared with attend-direction conditions. The other class displayed the opposite effects. Feature-selective activity modulations occurred earlier in neurons modulated by attention to color compared with neurons modulated by attention to motion direction. Thus feature-selective attention influences neuronal processing in macaque area MT but often exhibited a mismatch between the preferred stimulus dimension (direction of motion) and the preferred attention dimension (attention to color). PMID:22170961

Effect of feature-selective attention on neuronal responses in macaque area MT.

PubMed

Chen, X; Hoffmann, K-P; Albright, T D; Thiele, A

2012-03-01

Attention influences visual processing in striate and extrastriate cortex, which has been extensively studied for spatial-, object-, and feature-based attention. Most studies exploring neural signatures of feature-based attention have trained animals to attend to an object identified by a certain feature and ignore objects/displays identified by a different feature. Little is known about the effects of feature-selective attention, where subjects attend to one stimulus feature domain (e.g., color) of an object while features from different domains (e.g., direction of motion) of the same object are ignored. To study this type of feature-selective attention in area MT in the middle temporal sulcus, we trained macaque monkeys to either attend to and report the direction of motion of a moving sine wave grating (a feature for which MT neurons display strong selectivity) or attend to and report its color (a feature for which MT neurons have very limited selectivity). We hypothesized that neurons would upregulate their firing rate during attend-direction conditions compared with attend-color conditions. We found that feature-selective attention significantly affected 22% of MT neurons. Contrary to our hypothesis, these neurons did not necessarily increase firing rate when animals attended to direction of motion but fell into one of two classes. In one class, attention to color increased the gain of stimulus-induced responses compared with attend-direction conditions. The other class displayed the opposite effects. Feature-selective activity modulations occurred earlier in neurons modulated by attention to color compared with neurons modulated by attention to motion direction. Thus feature-selective attention influences neuronal processing in macaque area MT but often exhibited a mismatch between the preferred stimulus dimension (direction of motion) and the preferred attention dimension (attention to color).

Estimation of pyrethroid pesticide intake using regression modeling of food groups based on composite dietary samples

EPA Science Inventory

Population-based estimates of pesticide intake are needed to characterize exposure for particular demographic groups based on their dietary behaviors. Regression modeling performed on measurements of selected pesticides in composited duplicate diet samples allowed (1) estimation ...

Estimation of pyrethroid pesticide intake using regression modeling of food groups based on composite dietary samples..

EPA Science Inventory

Population-based estimates of pesticide intake are needed to characterize exposure for particular demographic groups based on their dietary behaviors. Regression modeling performed on measurements of selected pesticides in composited duplicate diet samples allowed (1) estimation ...

Estimation of pyrethroid pesticide intake using regression modeling of food groups based on composite dietary samples.

EPA Science Inventory

Population-based estimates of pesticide intake are needed to characterize exposure for particular demographic groups based on their dietary behaviors. Regression modeling performed on measurements of selected pesticides in composited duplicate diet samples allowed (1) estimation ...

An improved wrapper-based feature selection method for machinery fault diagnosis

PubMed Central

2017-01-01

A major issue of machinery fault diagnosis using vibration signals is that it is over-reliant on personnel knowledge and experience in interpreting the signal. Thus, machine learning has been adapted for machinery fault diagnosis. The quantity and quality of the input features, however, influence the fault classification performance. Feature selection plays a vital role in selecting the most representative feature subset for the machine learning algorithm. In contrast, the trade-off relationship between capability when selecting the best feature subset and computational effort is inevitable in the wrapper-based feature selection (WFS) method. This paper proposes an improved WFS technique before integration with a support vector machine (SVM) model classifier as a complete fault diagnosis system for a rolling element bearing case study. The bearing vibration dataset made available by the Case Western Reserve University Bearing Data Centre was executed using the proposed WFS and its performance has been analysed and discussed. The results reveal that the proposed WFS secures the best feature subset with a lower computational effort by eliminating the redundancy of re-evaluation. The proposed WFS has therefore been found to be capable and efficient to carry out feature selection tasks. PMID:29261689

[Optimization of diagnosis indicator selection and inspection plan by 3.0T MRI in breast cancer].

PubMed

Jiang, Zhongbiao; Wang, Yunhua; He, Zhong; Zhang, Lejun; Zheng, Kai

2013-08-01

To optimize 3.0T MRI diagnosis indicator in breast cancer and to select the best MRI scan program. Totally 45 patients with breast cancers were collected, and another 35 patients with benign breast tumor served as the control group. All patients underwent 3.0T MRI, including T1- weighted imaging (T1WI), fat suppression of the T2-weighted imaging (T2WI), diffusion weighted imaging (DWI), 1H magnetic resonance spectroscopy (1H-MRS) and dynamic contrast enhanced (DCE) sequence. With operation pathology results as the gold standard in the diagnosis of breast diseases, the pathological results of benign and malignant served as dependent variables, and the diagnostic indicators of MRI were taken as independent variables. We put all the indicators of MRI examination under Logistic regression analysis, established the Logistic model, and optimized the diagnosis indicators of MRI examination to further improve MRI scan of breast cancer. By Logistic regression analysis, some indicators were selected in the equation, including the edge feature of the tumor, the time-signal intensity curve (TIC) type and the apparent diffusion coefficient (ADC) value when b=500 s/mm2. The regression equation was Logit (P)=-21.936+20.478X6+3.267X7+ 21.488X3. Valuable indicators in the diagnosis of breast cancer are the edge feature of the tumor, the TIC type and the ADC value when b=500 s/mm2. Combining conventional MRI scan, DWI and dynamic enhanced MRI is a better examination program, while MRS is the complementary program when diagnosis is difficult.

Detrended fluctuation analysis as a regression framework: Estimating dependence at different scales

NASA Astrophysics Data System (ADS)

Kristoufek, Ladislav

2015-02-01

We propose a framework combining detrended fluctuation analysis with standard regression methodology. The method is built on detrended variances and covariances and it is designed to estimate regression parameters at different scales and under potential nonstationarity and power-law correlations. The former feature allows for distinguishing between effects for a pair of variables from different temporal perspectives. The latter ones make the method a significant improvement over the standard least squares estimation. Theoretical claims are supported by Monte Carlo simulations. The method is then applied on selected examples from physics, finance, environmental science, and epidemiology. For most of the studied cases, the relationship between variables of interest varies strongly across scales.

Improved sparse decomposition based on a smoothed L0 norm using a Laplacian kernel to select features from fMRI data.

PubMed

Zhang, Chuncheng; Song, Sutao; Wen, Xiaotong; Yao, Li; Long, Zhiying

2015-04-30

Feature selection plays an important role in improving the classification accuracy of multivariate classification techniques in the context of fMRI-based decoding due to the "few samples and large features" nature of functional magnetic resonance imaging (fMRI) data. Recently, several sparse representation methods have been applied to the voxel selection of fMRI data. Despite the low computational efficiency of the sparse representation methods, they still displayed promise for applications that select features from fMRI data. In this study, we proposed the Laplacian smoothed L0 norm (LSL0) approach for feature selection of fMRI data. Based on the fast sparse decomposition using smoothed L0 norm (SL0) (Mohimani, 2007), the LSL0 method used the Laplacian function to approximate the L0 norm of sources. Results of the simulated and real fMRI data demonstrated the feasibility and robustness of LSL0 for the sparse source estimation and feature selection. Simulated results indicated that LSL0 produced more accurate source estimation than SL0 at high noise levels. The classification accuracy using voxels that were selected by LSL0 was higher than that by SL0 in both simulated and real fMRI experiment. Moreover, both LSL0 and SL0 showed higher classification accuracy and required less time than ICA and t-test for the fMRI decoding. LSL0 outperformed SL0 in sparse source estimation at high noise level and in feature selection. Moreover, LSL0 and SL0 showed better performance than ICA and t-test for feature selection. Copyright © 2015 Elsevier B.V. All rights reserved.

On the use of feature selection to improve the detection of sea oil spills in SAR images

NASA Astrophysics Data System (ADS)

Mera, David; Bolon-Canedo, Veronica; Cotos, J. M.; Alonso-Betanzos, Amparo

2017-03-01

Fast and effective oil spill detection systems are crucial to ensure a proper response to environmental emergencies caused by hydrocarbon pollution on the ocean's surface. Typically, these systems uncover not only oil spills, but also a high number of look-alikes. The feature extraction is a critical and computationally intensive phase where each detected dark spot is independently examined. Traditionally, detection systems use an arbitrary set of features to discriminate between oil spills and look-alikes phenomena. However, Feature Selection (FS) methods based on Machine Learning (ML) have proved to be very useful in real domains for enhancing the generalization capabilities of the classifiers, while discarding the existing irrelevant features. In this work, we present a generic and systematic approach, based on FS methods, for choosing a concise and relevant set of features to improve the oil spill detection systems. We have compared five FS methods: Correlation-based feature selection (CFS), Consistency-based filter, Information Gain, ReliefF and Recursive Feature Elimination for Support Vector Machine (SVM-RFE). They were applied on a 141-input vector composed of features from a collection of outstanding studies. Selected features were validated via a Support Vector Machine (SVM) classifier and the results were compared with previous works. Test experiments revealed that the classifier trained with the 6-input feature vector proposed by SVM-RFE achieved the best accuracy and Cohen's kappa coefficient (87.1% and 74.06% respectively). This is a smaller feature combination with similar or even better classification accuracy than previous works. The presented finding allows to speed up the feature extraction phase without reducing the classifier accuracy. Experiments also confirmed the significance of the geometrical features since 75.0% of the different features selected by the applied FS methods as well as 66.67% of the proposed 6-input feature vector belong to this category.

Integration of Error Compensation of Coordinate Measuring Machines into Feature Measurement: Part II—Experimental Implementation

PubMed Central

Calvo, Roque; D’Amato, Roberto; Gómez, Emilio; Domingo, Rosario

2016-01-01

Coordinate measuring machines (CMM) are main instruments of measurement in laboratories and in industrial quality control. A compensation error model has been formulated (Part I). It integrates error and uncertainty in the feature measurement model. Experimental implementation for the verification of this model is carried out based on the direct testing on a moving bridge CMM. The regression results by axis are quantified and compared to CMM indication with respect to the assigned values of the measurand. Next, testing of selected measurements of length, flatness, dihedral angle, and roundness features are accomplished. The measurement of calibrated gauge blocks for length or angle, flatness verification of the CMM granite table and roundness of a precision glass hemisphere are presented under a setup of repeatability conditions. The results are analysed and compared with alternative methods of estimation. The overall performance of the model is endorsed through experimental verification, as well as the practical use and the model capability to contribute in the improvement of current standard CMM measuring capabilities. PMID:27754441

In silico quantitative structure-toxicity relationship study of aromatic nitro compounds.

PubMed

Pasha, Farhan Ahmad; Neaz, Mohammad Morshed; Cho, Seung Joo; Ansari, Mohiuddin; Mishra, Sunil Kumar; Tiwari, Sharvan

2009-05-01

Small molecules often have toxicities that are a function of molecular structural features. Minor variations in structural features can make large difference in such toxicity. Consequently, in silico techniques may be used to correlate such molecular toxicities with their structural features. Relative to nine different sets of aromatic nitro compounds having known observed toxicities against different targets, we developed ligand-based 2D quantitative structure-toxicity relationship models using 20 selected topological descriptors. The topological descriptors have several advantages such as conformational independency, facile and less time-consuming computation to yield good results. Multiple linear regression analysis was used to correlate variations of toxicity with molecular properties. The information index on molecular size, lopping centric index and Kier flexibility index were identified as fundamental descriptors for different kinds of toxicity, and further showed that molecular size, branching and molecular flexibility might be particularly important factors in quantitative structure-toxicity relationship analysis. This study revealed that topological descriptor-guided quantitative structure-toxicity relationship provided a very useful, cost and time-efficient, in silico tool for describing small-molecule toxicities.

EDA-gram: designing electrodermal activity fingerprints for visualization and feature extraction.

PubMed

Chaspari, Theodora; Tsiartas, Andreas; Stein Duker, Leah I; Cermak, Sharon A; Narayanan, Shrikanth S

2016-08-01

Wearable technology permeates every aspect of our daily life increasing the need of reliable and interpretable models for processing the large amount of biomedical data. We propose the EDA-Gram, a multidimensional fingerprint of the electrodermal activity (EDA) signal, inspired by the widely-used notion of spectrogram. The EDA-Gram is based on the sparse decomposition of EDA from a knowledge-driven set of dictionary atoms. The time axis reflects the analysis frames, the spectral dimension depicts the width of selected dictionary atoms, while intensity values are computed from the atom coefficients. In this way, EDA-Gram incorporates the amplitude and shape of Skin Conductance Responses (SCR), which comprise an essential part of the signal. EDA-Gram is further used as a foundation for signal-specific feature design. Our results indicate that the proposed representation can accentuate fine-grain signal fluctuations, which might not always be apparent through simple visual inspection. Statistical analysis and classification/regression experiments further suggest that the derived features can differentiate between multiple arousal levels and stress-eliciting environments for two datasets.

Optimal subset selection of primary sequence features using the genetic algorithm for thermophilic proteins identification.

PubMed

Wang, LiQiang; Li, CuiFeng

2014-10-01

A genetic algorithm (GA) coupled with multiple linear regression (MLR) was used to extract useful features from amino acids and g-gap dipeptides for distinguishing between thermophilic and non-thermophilic proteins. The method was trained by a benchmark dataset of 915 thermophilic and 793 non-thermophilic proteins. The method reached an overall accuracy of 95.4 % in a Jackknife test using nine amino acids, 38 0-gap dipeptides and 29 1-gap dipeptides. The accuracy as a function of protein size ranged between 85.8 and 96.9 %. The overall accuracies of three independent tests were 93, 93.4 and 91.8 %. The observed results of detecting thermophilic proteins suggest that the GA-MLR approach described herein should be a powerful method for selecting features that describe thermostabile machines and be an aid in the design of more stable proteins.

Binding Affinity prediction with Property Encoded Shape Distribution signatures

PubMed Central

Das, Sourav; Krein, Michael P.

2010-01-01

We report the use of the molecular signatures known as “Property-Encoded Shape Distributions” (PESD) together with standard Support Vector Machine (SVM) techniques to produce validated models that can predict the binding affinity of a large number of protein ligand complexes. This “PESD-SVM” method uses PESD signatures that encode molecular shapes and property distributions on protein and ligand surfaces as features to build SVM models that require no subjective feature selection. A simple protocol was employed for tuning the SVM models during their development, and the results were compared to SFCscore – a regression-based method that was previously shown to perform better than 14 other scoring functions. Although the PESD-SVM method is based on only two surface property maps, the overall results were comparable. For most complexes with a dominant enthalpic contribution to binding (ΔH/-TΔS > 3), a good correlation between true and predicted affinities was observed. Entropy and solvent were not considered in the present approach and further improvement in accuracy would require accounting for these components rigorously. PMID:20095526

Feature Augmentation via Nonparametrics and Selection (FANS) in High-Dimensional Classification.

PubMed

Fan, Jianqing; Feng, Yang; Jiang, Jiancheng; Tong, Xin

We propose a high dimensional classification method that involves nonparametric feature augmentation. Knowing that marginal density ratios are the most powerful univariate classifiers, we use the ratio estimates to transform the original feature measurements. Subsequently, penalized logistic regression is invoked, taking as input the newly transformed or augmented features. This procedure trains models equipped with local complexity and global simplicity, thereby avoiding the curse of dimensionality while creating a flexible nonlinear decision boundary. The resulting method is called Feature Augmentation via Nonparametrics and Selection (FANS). We motivate FANS by generalizing the Naive Bayes model, writing the log ratio of joint densities as a linear combination of those of marginal densities. It is related to generalized additive models, but has better interpretability and computability. Risk bounds are developed for FANS. In numerical analysis, FANS is compared with competing methods, so as to provide a guideline on its best application domain. Real data analysis demonstrates that FANS performs very competitively on benchmark email spam and gene expression data sets. Moreover, FANS is implemented by an extremely fast algorithm through parallel computing.

Feature Augmentation via Nonparametrics and Selection (FANS) in High-Dimensional Classification

PubMed Central

Feng, Yang; Jiang, Jiancheng; Tong, Xin

2015-01-01

We propose a high dimensional classification method that involves nonparametric feature augmentation. Knowing that marginal density ratios are the most powerful univariate classifiers, we use the ratio estimates to transform the original feature measurements. Subsequently, penalized logistic regression is invoked, taking as input the newly transformed or augmented features. This procedure trains models equipped with local complexity and global simplicity, thereby avoiding the curse of dimensionality while creating a flexible nonlinear decision boundary. The resulting method is called Feature Augmentation via Nonparametrics and Selection (FANS). We motivate FANS by generalizing the Naive Bayes model, writing the log ratio of joint densities as a linear combination of those of marginal densities. It is related to generalized additive models, but has better interpretability and computability. Risk bounds are developed for FANS. In numerical analysis, FANS is compared with competing methods, so as to provide a guideline on its best application domain. Real data analysis demonstrates that FANS performs very competitively on benchmark email spam and gene expression data sets. Moreover, FANS is implemented by an extremely fast algorithm through parallel computing. PMID:27185970

Effectively identifying compound-protein interactions by learning from positive and unlabeled examples.

PubMed

Cheng, Zhanzhan; Zhou, Shuigeng; Wang, Yang; Liu, Hui; Guan, Jihong; Chen, Yi-Ping Phoebe

2016-05-18

Prediction of compound-protein interactions (CPIs) is to find new compound-protein pairs where a protein is targeted by at least a compound, which is a crucial step in new drug design. Currently, a number of machine learning based methods have been developed to predict new CPIs in the literature. However, as there is not yet any publicly available set of validated negative CPIs, most existing machine learning based approaches use the unknown interactions (not validated CPIs) selected randomly as the negative examples to train classifiers for predicting new CPIs. Obviously, this is not quite reasonable and unavoidably impacts the CPI prediction performance. In this paper, we simply take the unknown CPIs as unlabeled examples, and propose a new method called PUCPI (the abbreviation of PU learning for Compound-Protein Interaction identification) that employs biased-SVM (Support Vector Machine) to predict CPIs using only positive and unlabeled examples. PU learning is a class of learning methods that leans from positive and unlabeled (PU) samples. To the best of our knowledge, this is the first work that identifies CPIs using only positive and unlabeled examples. We first collect known CPIs as positive examples and then randomly select compound-protein pairs not in the positive set as unlabeled examples. For each CPI/compound-protein pair, we extract protein domains as protein features and compound substructures as chemical features, then take the tensor product of the corresponding compound features and protein features as the feature vector of the CPI/compound-protein pair. After that, biased-SVM is employed to train classifiers on different datasets of CPIs and compound-protein pairs. Experiments over various datasets show that our method outperforms six typical classifiers, including random forest, L1- and L2-regularized logistic regression, naive Bayes, SVM and k-nearest neighbor (kNN), and three types of existing CPI prediction models. Source code, datasets and related documents of PUCPI are available at: http://admis.fudan.edu.cn/projects/pucpi.html.

Multiband tangent space mapping and feature selection for classification of EEG during motor imagery.

PubMed

Islam, Md Rabiul; Tanaka, Toshihisa; Molla, Md Khademul Islam

2018-05-08

When designing multiclass motor imagery-based brain-computer interface (MI-BCI), a so-called tangent space mapping (TSM) method utilizing the geometric structure of covariance matrices is an effective technique. This paper aims to introduce a method using TSM for finding accurate operational frequency bands related brain activities associated with MI tasks. A multichannel electroencephalogram (EEG) signal is decomposed into multiple subbands, and tangent features are then estimated on each subband. A mutual information analysis-based effective algorithm is implemented to select subbands containing features capable of improving motor imagery classification accuracy. Thus obtained features of selected subbands are combined to get feature space. A principal component analysis-based approach is employed to reduce the features dimension and then the classification is accomplished by a support vector machine (SVM). Offline analysis demonstrates the proposed multiband tangent space mapping with subband selection (MTSMS) approach outperforms state-of-the-art methods. It acheives the highest average classification accuracy for all datasets (BCI competition dataset 2a, IIIa, IIIb, and dataset JK-HH1). The increased classification accuracy of MI tasks with the proposed MTSMS approach can yield effective implementation of BCI. The mutual information-based subband selection method is implemented to tune operation frequency bands to represent actual motor imagery tasks.

Vessel Classification in Cosmo-Skymed SAR Data Using Hierarchical Feature Selection

NASA Astrophysics Data System (ADS)

Makedonas, A.; Theoharatos, C.; Tsagaris, V.; Anastasopoulos, V.; Costicoglou, S.

2015-04-01

SAR based ship detection and classification are important elements of maritime monitoring applications. Recently, high-resolution SAR data have opened new possibilities to researchers for achieving improved classification results. In this work, a hierarchical vessel classification procedure is presented based on a robust feature extraction and selection scheme that utilizes scale, shape and texture features in a hierarchical way. Initially, different types of feature extraction algorithms are implemented in order to form the utilized feature pool, able to represent the structure, material, orientation and other vessel type characteristics. A two-stage hierarchical feature selection algorithm is utilized next in order to be able to discriminate effectively civilian vessels into three distinct types, in COSMO-SkyMed SAR images: cargos, small ships and tankers. In our analysis, scale and shape features are utilized in order to discriminate smaller types of vessels present in the available SAR data, or shape specific vessels. Then, the most informative texture and intensity features are incorporated in order to be able to better distinguish the civilian types with high accuracy. A feature selection procedure that utilizes heuristic measures based on features' statistical characteristics, followed by an exhaustive research with feature sets formed by the most qualified features is carried out, in order to discriminate the most appropriate combination of features for the final classification. In our analysis, five COSMO-SkyMed SAR data with 2.2m x 2.2m resolution were used to analyse the detailed characteristics of these types of ships. A total of 111 ships with available AIS data were used in the classification process. The experimental results show that this method has good performance in ship classification, with an overall accuracy reaching 83%. Further investigation of additional features and proper feature selection is currently in progress.

Machine Learning Algorithms for prediction of regions of high Reynolds Averaged Navier Stokes Uncertainty

NASA Astrophysics Data System (ADS)

Mishra, Aashwin; Iaccarino, Gianluca

2017-11-01

In spite of their deficiencies, RANS models represent the workhorse for industrial investigations into turbulent flows. In this context, it is essential to provide diagnostic measures to assess the quality of RANS predictions. To this end, the primary step is to identify feature importances amongst massive sets of potentially descriptive and discriminative flow features. This aids the physical interpretability of the resultant discrepancy model and its extensibility to similar problems. Recent investigations have utilized approaches such as Random Forests, Support Vector Machines and the Least Absolute Shrinkage and Selection Operator for feature selection. With examples, we exhibit how such methods may not be suitable for turbulent flow datasets. The underlying rationale, such as the correlation bias and the required conditions for the success of penalized algorithms, are discussed with illustrative examples. Finally, we provide alternate approaches using convex combinations of regularized regression approaches and randomized sub-sampling in combination with feature selection algorithms, to infer model structure from data. This research was supported by the Defense Advanced Research Projects Agency under the Enabling Quantification of Uncertainty in Physical Systems (EQUiPS) project (technical monitor: Dr Fariba Fahroo).

Two Paradoxes in Linear Regression Analysis.

PubMed

Feng, Ge; Peng, Jing; Tu, Dongke; Zheng, Julia Z; Feng, Changyong

2016-12-25

Regression is one of the favorite tools in applied statistics. However, misuse and misinterpretation of results from regression analysis are common in biomedical research. In this paper we use statistical theory and simulation studies to clarify some paradoxes around this popular statistical method. In particular, we show that a widely used model selection procedure employed in many publications in top medical journals is wrong. Formal procedures based on solid statistical theory should be used in model selection.

A Machine Learning Approach to Measurement of Text Readability for EFL Learners Using Various Linguistic Features

ERIC Educational Resources Information Center

Kotani, Katsunori; Yoshimi, Takehiko; Isahara, Hitoshi

2011-01-01

The present paper introduces and evaluates a readability measurement method designed for learners of EFL (English as a foreign language). The proposed readability measurement method (a regression model) estimates the text readability based on linguistic features, such as lexical, syntactic and discourse features. Text readability refers to the…

Analysis of the quality of image data acquired by the LANDSAT-4 Thematic Mapper and Multispectral Scanners

NASA Technical Reports Server (NTRS)

Colwell, R. N. (Principal Investigator)

1984-01-01

The geometric quality of TM film and digital products is evaluated by making selective photomeasurements and by measuring the coordinates of known features on both the TM products and map products. These paired observations are related using a standard linear least squares regression approach. Using regression equations and coefficients developed from 225 (TM film product) and 20 (TM digital product) control points, map coordinates of test points are predicted. The residual error vectors and analysis of variance (ANOVA) were performed on the east and north residual using nine image segments (blocks) as treatments. Based on the root mean square error of the 223 (TM film product) and 22 (TM digital product) test points, users of TM data expect the planimetric accuracy of mapped points to be within 91 meters and within 117 meters for the film products, and to be within 12 meters and within 14 meters for the digital products.

Computerized detection of breast cancer using resonance-frequency-based electrical impedance spectroscopy

NASA Astrophysics Data System (ADS)

Gao, Wei; Fan, Ming; Zhao, Weijie; Zheng, Bin; Li, Lihua

2017-03-01

This study developed and tested a multi-probe resonance-frequency-based electrical impedance spectroscopy (REIS) system aimed at detection of breast cancer. The REIS system consists of specially designed mechanical supporting device that can be easily lifted to fit women of different height, a seven probe sensor cup, and a computer providing software for system control and management. The sensor cup includes one central probe for direct contact with the nipple, and other six probes uniformly distributed at a distance of 35mm away from the center probe to enable contact with breast skin surface. It takes about 18 seconds for this system to complete a data acquisition process. We utilized this system for examination of breast cancer, collecting a dataset of 289 cases including biopsy verified 74 malignant and 215 benign tumors. After that, 23 REIS based features, including seven frequency, fifteen magnitude features were extracted, and an age feature. To reduce redundancy we selected 6 features using the evolutionary algorithm for classification. The area under a receiver operating characteristic curve (AUC) was computed to assess classifier performance. A multivariable logistic regression method was performed for detection of the tumors. The results of our study showed for the 23 REIS features AUC and ACC, Sensitivity and Specificity of 0.796, 0.727, 0.731 and 0.726, respectively. The AUC and ACC, Sensitivity and Specificity for the 6 REIS features of 0.840, 0.80, 0.703 and 0.833, respectively, and AUC of 0.662 and 0.619 for the frequency and magnitude based REIS features, respectively. The performance of the classifiers using all the 6 features was significantly better than solely using magnitude features (p=3.29e-08) and frequency features (5.61e-07). Smote algorithm was used to expand small samples to balance the dataset, the AUC after data balance of 0.846 increased than the original data classification performance. The results indicated that the REIS system is a promising tool for detection of breast cancer and may be acceptable for clinical implementation.

[Electroencephalogram Feature Selection Based on Correlation Coefficient Analysis].

PubMed

Zhou, Jinzhi; Tang, Xiaofang

2015-08-01

In order to improve the accuracy of classification with small amount of motor imagery training data on the development of brain-computer interface (BCD systems, we proposed an analyzing method to automatically select the characteristic parameters based on correlation coefficient analysis. Throughout the five sample data of dataset IV a from 2005 BCI Competition, we utilized short-time Fourier transform (STFT) and correlation coefficient calculation to reduce the number of primitive electroencephalogram dimension, then introduced feature extraction based on common spatial pattern (CSP) and classified by linear discriminant analysis (LDA). Simulation results showed that the average rate of classification accuracy could be improved by using correlation coefficient feature selection method than those without using this algorithm. Comparing with support vector machine (SVM) optimization features algorithm, the correlation coefficient analysis can lead better selection parameters to improve the accuracy of classification.

A new computational strategy for predicting essential genes.

PubMed

Cheng, Jian; Wu, Wenwu; Zhang, Yinwen; Li, Xiangchen; Jiang, Xiaoqian; Wei, Gehong; Tao, Shiheng

2013-12-21

Determination of the minimum gene set for cellular life is one of the central goals in biology. Genome-wide essential gene identification has progressed rapidly in certain bacterial species; however, it remains difficult to achieve in most eukaryotic species. Several computational models have recently been developed to integrate gene features and used as alternatives to transfer gene essentiality annotations between organisms. We first collected features that were widely used by previous predictive models and assessed the relationships between gene features and gene essentiality using a stepwise regression model. We found two issues that could significantly reduce model accuracy: (i) the effect of multicollinearity among gene features and (ii) the diverse and even contrasting correlations between gene features and gene essentiality existing within and among different species. To address these issues, we developed a novel model called feature-based weighted Naïve Bayes model (FWM), which is based on Naïve Bayes classifiers, logistic regression, and genetic algorithm. The proposed model assesses features and filters out the effects of multicollinearity and diversity. The performance of FWM was compared with other popular models, such as support vector machine, Naïve Bayes model, and logistic regression model, by applying FWM to reciprocally predict essential genes among and within 21 species. Our results showed that FWM significantly improves the accuracy and robustness of essential gene prediction. FWM can remarkably improve the accuracy of essential gene prediction and may be used as an alternative method for other classification work. This method can contribute substantially to the knowledge of the minimum gene sets required for living organisms and the discovery of new drug targets.

Neural mechanisms of selective attention in the somatosensory system.

PubMed

Gomez-Ramirez, Manuel; Hysaj, Kristjana; Niebur, Ernst

2016-09-01

Selective attention allows organisms to extract behaviorally relevant information while ignoring distracting stimuli that compete for the limited resources of their central nervous systems. Attention is highly flexible, and it can be harnessed to select information based on sensory modality, within-modality feature(s), spatial location, object identity, and/or temporal properties. In this review, we discuss the body of work devoted to understanding mechanisms of selective attention in the somatosensory system. In particular, we describe the effects of attention on tactile behavior and corresponding neural activity in somatosensory cortex. Our focus is on neural mechanisms that select tactile stimuli based on their location on the body (somatotopic-based attention) or their sensory feature (feature-based attention). We highlight parallels between selection mechanisms in touch and other sensory systems and discuss several putative neural coding schemes employed by cortical populations to signal the behavioral relevance of sensory inputs. Specifically, we contrast the advantages and disadvantages of using a gain vs. spike-spike correlation code for representing attended sensory stimuli. We favor a neural network model of tactile attention that is composed of frontal, parietal, and subcortical areas that controls somatosensory cells encoding the relevant stimulus features to enable preferential processing throughout the somatosensory hierarchy. Our review is based on data from noninvasive electrophysiological and imaging data in humans as well as single-unit recordings in nonhuman primates. Copyright © 2016 the American Physiological Society.

Neural mechanisms of selective attention in the somatosensory system

PubMed Central

Hysaj, Kristjana; Niebur, Ernst

2016-01-01

Selective attention allows organisms to extract behaviorally relevant information while ignoring distracting stimuli that compete for the limited resources of their central nervous systems. Attention is highly flexible, and it can be harnessed to select information based on sensory modality, within-modality feature(s), spatial location, object identity, and/or temporal properties. In this review, we discuss the body of work devoted to understanding mechanisms of selective attention in the somatosensory system. In particular, we describe the effects of attention on tactile behavior and corresponding neural activity in somatosensory cortex. Our focus is on neural mechanisms that select tactile stimuli based on their location on the body (somatotopic-based attention) or their sensory feature (feature-based attention). We highlight parallels between selection mechanisms in touch and other sensory systems and discuss several putative neural coding schemes employed by cortical populations to signal the behavioral relevance of sensory inputs. Specifically, we contrast the advantages and disadvantages of using a gain vs. spike-spike correlation code for representing attended sensory stimuli. We favor a neural network model of tactile attention that is composed of frontal, parietal, and subcortical areas that controls somatosensory cells encoding the relevant stimulus features to enable preferential processing throughout the somatosensory hierarchy. Our review is based on data from noninvasive electrophysiological and imaging data in humans as well as single-unit recordings in nonhuman primates. PMID:27334956

Patient Stratification Using Electronic Health Records from a Chronic Disease Management Program.

PubMed

Chen, Robert; Sun, Jimeng; Dittus, Robert S; Fabbri, Daniel; Kirby, Jacqueline; Laffer, Cheryl L; McNaughton, Candace D; Malin, Bradley

2016-01-04

The goal of this study is to devise a machine learning framework to assist care coordination programs in prognostic stratification to design and deliver personalized care plans and to allocate financial and medical resources effectively. This study is based on a de-identified cohort of 2,521 hypertension patients from a chronic care coordination program at the Vanderbilt University Medical Center. Patients were modeled as vectors of features derived from electronic health records (EHRs) over a six-year period. We applied a stepwise regression to identify risk factors associated with a decrease in mean arterial pressure of at least 2 mmHg after program enrollment. The resulting features were subsequently validated via a logistic regression classifier. Finally, risk factors were applied to group the patients through model-based clustering. We identified a set of predictive features that consisted of a mix of demographic, medication, and diagnostic concepts. Logistic regression over these features yielded an area under the ROC curve (AUC) of 0.71 (95% CI: [0.67, 0.76]). Based on these features, four clinically meaningful groups are identified through clustering - two of which represented patients with more severe disease profiles, while the remaining represented patients with mild disease profiles. Patients with hypertension can exhibit significant variation in their blood pressure control status and responsiveness to therapy. Yet this work shows that a clustering analysis can generate more homogeneous patient groups, which may aid clinicians in designing and implementing customized care programs. The study shows that predictive modeling and clustering using EHR data can be beneficial for providing a systematic, generalized approach for care providers to tailor their management approach based upon patient-level factors.

Sparse Zero-Sum Games as Stable Functional Feature Selection

PubMed Central

Sokolovska, Nataliya; Teytaud, Olivier; Rizkalla, Salwa; Clément, Karine; Zucker, Jean-Daniel

2015-01-01

In large-scale systems biology applications, features are structured in hidden functional categories whose predictive power is identical. Feature selection, therefore, can lead not only to a problem with a reduced dimensionality, but also reveal some knowledge on functional classes of variables. In this contribution, we propose a framework based on a sparse zero-sum game which performs a stable functional feature selection. In particular, the approach is based on feature subsets ranking by a thresholding stochastic bandit. We provide a theoretical analysis of the introduced algorithm. We illustrate by experiments on both synthetic and real complex data that the proposed method is competitive from the predictive and stability viewpoints. PMID:26325268

An ant colony optimization based feature selection for web page classification.

PubMed

Saraç, Esra; Özel, Selma Ayşe

2014-01-01

The increased popularity of the web has caused the inclusion of huge amount of information to the web, and as a result of this explosive information growth, automated web page classification systems are needed to improve search engines' performance. Web pages have a large number of features such as HTML/XML tags, URLs, hyperlinks, and text contents that should be considered during an automated classification process. The aim of this study is to reduce the number of features to be used to improve runtime and accuracy of the classification of web pages. In this study, we used an ant colony optimization (ACO) algorithm to select the best features, and then we applied the well-known C4.5, naive Bayes, and k nearest neighbor classifiers to assign class labels to web pages. We used the WebKB and Conference datasets in our experiments, and we showed that using the ACO for feature selection improves both accuracy and runtime performance of classification. We also showed that the proposed ACO based algorithm can select better features with respect to the well-known information gain and chi square feature selection methods.

A machine learning-based framework to identify type 2 diabetes through electronic health records

PubMed Central

Zheng, Tao; Xie, Wei; Xu, Liling; He, Xiaoying; Zhang, Ya; You, Mingrong; Yang, Gong; Chen, You

2016-01-01

Objective To discover diverse genotype-phenotype associations affiliated with Type 2 Diabetes Mellitus (T2DM) via genome-wide association study (GWAS) and phenome-wide association study (PheWAS), more cases (T2DM subjects) and controls (subjects without T2DM) are required to be identified (e.g., via Electronic Health Records (EHR)). However, existing expert based identification algorithms often suffer in a low recall rate and could miss a large number of valuable samples under conservative filtering standards. The goal of this work is to develop a semi-automated framework based on machine learning as a pilot study to liberalize filtering criteria to improve recall rate with a keeping of low false positive rate. Materials and methods We propose a data informed framework for identifying subjects with and without T2DM from EHR via feature engineering and machine learning. We evaluate and contrast the identification performance of widely-used machine learning models within our framework, including k-Nearest-Neighbors, Naïve Bayes, Decision Tree, Random Forest, Support Vector Machine and Logistic Regression. Our framework was conducted on 300 patient samples (161 cases, 60 controls and 79 unconfirmed subjects), randomly selected from 23,281 diabetes related cohort retrieved from a regional distributed EHR repository ranging from 2012 to 2014. Results We apply top-performing machine learning algorithms on the engineered features. We benchmark and contrast the accuracy, precision, AUC, sensitivity and specificity of classification models against the state-of-the-art expert algorithm for identification of T2DM subjects. Our results indicate that the framework achieved high identification performances (∼0.98 in average AUC), which are much higher than the state-of-the-art algorithm (0.71 in AUC). Discussion Expert algorithm-based identification of T2DM subjects from EHR is often hampered by the high missing rates due to their conservative selection criteria. Our framework leverages machine learning and feature engineering to loosen such selection criteria to achieve a high identification rate of cases and controls. Conclusions Our proposed framework demonstrates a more accurate and efficient approach for identifying subjects with and without T2DM from EHR. PMID:27919371

A machine learning-based framework to identify type 2 diabetes through electronic health records.

PubMed

Zheng, Tao; Xie, Wei; Xu, Liling; He, Xiaoying; Zhang, Ya; You, Mingrong; Yang, Gong; Chen, You

2017-01-01

To discover diverse genotype-phenotype associations affiliated with Type 2 Diabetes Mellitus (T2DM) via genome-wide association study (GWAS) and phenome-wide association study (PheWAS), more cases (T2DM subjects) and controls (subjects without T2DM) are required to be identified (e.g., via Electronic Health Records (EHR)). However, existing expert based identification algorithms often suffer in a low recall rate and could miss a large number of valuable samples under conservative filtering standards. The goal of this work is to develop a semi-automated framework based on machine learning as a pilot study to liberalize filtering criteria to improve recall rate with a keeping of low false positive rate. We propose a data informed framework for identifying subjects with and without T2DM from EHR via feature engineering and machine learning. We evaluate and contrast the identification performance of widely-used machine learning models within our framework, including k-Nearest-Neighbors, Naïve Bayes, Decision Tree, Random Forest, Support Vector Machine and Logistic Regression. Our framework was conducted on 300 patient samples (161 cases, 60 controls and 79 unconfirmed subjects), randomly selected from 23,281 diabetes related cohort retrieved from a regional distributed EHR repository ranging from 2012 to 2014. We apply top-performing machine learning algorithms on the engineered features. We benchmark and contrast the accuracy, precision, AUC, sensitivity and specificity of classification models against the state-of-the-art expert algorithm for identification of T2DM subjects. Our results indicate that the framework achieved high identification performances (∼0.98 in average AUC), which are much higher than the state-of-the-art algorithm (0.71 in AUC). Expert algorithm-based identification of T2DM subjects from EHR is often hampered by the high missing rates due to their conservative selection criteria. Our framework leverages machine learning and feature engineering to loosen such selection criteria to achieve a high identification rate of cases and controls. Our proposed framework demonstrates a more accurate and efficient approach for identifying subjects with and without T2DM from EHR. Copyright © 2016 Elsevier Ireland Ltd. All rights reserved.

Quantitative imaging features of pretreatment CT predict volumetric response to chemotherapy in patients with colorectal liver metastases.

PubMed

Creasy, John M; Midya, Abhishek; Chakraborty, Jayasree; Adams, Lauryn B; Gomes, Camilla; Gonen, Mithat; Seastedt, Kenneth P; Sutton, Elizabeth J; Cercek, Andrea; Kemeny, Nancy E; Shia, Jinru; Balachandran, Vinod P; Kingham, T Peter; Allen, Peter J; DeMatteo, Ronald P; Jarnagin, William R; D'Angelica, Michael I; Do, Richard K G; Simpson, Amber L

2018-06-19

This study investigates whether quantitative image analysis of pretreatment CT scans can predict volumetric response to chemotherapy for patients with colorectal liver metastases (CRLM). Patients treated with chemotherapy for CRLM (hepatic artery infusion (HAI) combined with systemic or systemic alone) were included in the study. Patients were imaged at baseline and approximately 8 weeks after treatment. Response was measured as the percentage change in tumour volume from baseline. Quantitative imaging features were derived from the index hepatic tumour on pretreatment CT, and features statistically significant on univariate analysis were included in a linear regression model to predict volumetric response. The regression model was constructed from 70% of data, while 30% were reserved for testing. Test data were input into the trained model. Model performance was evaluated with mean absolute prediction error (MAPE) and R 2 . Clinicopatholologic factors were assessed for correlation with response. 157 patients were included, split into training (n = 110) and validation (n = 47) sets. MAPE from the multivariate linear regression model was 16.5% (R 2 = 0.774) and 21.5% in the training and validation sets, respectively. Stratified by HAI utilisation, MAPE in the validation set was 19.6% for HAI and 25.1% for systemic chemotherapy alone. Clinical factors associated with differences in median tumour response were treatment strategy, systemic chemotherapy regimen, age and KRAS mutation status (p < 0.05). Quantitative imaging features extracted from pretreatment CT are promising predictors of volumetric response to chemotherapy in patients with CRLM. Pretreatment predictors of response have the potential to better select patients for specific therapies. • Colorectal liver metastases (CRLM) are downsized with chemotherapy but predicting the patients that will respond to chemotherapy is currently not possible. • Heterogeneity and enhancement patterns of CRLM can be measured with quantitative imaging. • Prediction model constructed that predicts volumetric response with 20% error suggesting that quantitative imaging holds promise to better select patients for specific treatments.

Analysis of the quality of image data acquired by the LANDSAT-4 thematic mapper and multispectral scanners

NASA Technical Reports Server (NTRS)

Colwell, R. N. (Principal Investigator)

1983-01-01

The geometric quality of the TM and MSS film products were evaluated by making selective photo measurements such as scale, linear and area determinations; and by measuring the coordinates of known features on both the film products and map products and then relating these paired observations using a standard linear least squares regression approach. Quantitative interpretation tests are described which evaluate the quality and utility of the TM film products and various band combinations for detecting and identifying important forest and agricultural features.

Diagnosing and ranking retinopathy disease level using diabetic fundus image recuperation approach.

PubMed

Somasundaram, K; Rajendran, P Alli

2015-01-01

Retinal fundus images are widely used in diagnosing different types of eye diseases. The existing methods such as Feature Based Macular Edema Detection (FMED) and Optimally Adjusted Morphological Operator (OAMO) effectively detected the presence of exudation in fundus images and identified the true positive ratio of exudates detection, respectively. These mechanically detected exudates did not include more detailed feature selection technique to the system for detection of diabetic retinopathy. To categorize the exudates, Diabetic Fundus Image Recuperation (DFIR) method based on sliding window approach is developed in this work to select the features of optic cup in digital retinal fundus images. The DFIR feature selection uses collection of sliding windows with varying range to obtain the features based on the histogram value using Group Sparsity Nonoverlapping Function. Using support vector model in the second phase, the DFIR method based on Spiral Basis Function effectively ranks the diabetic retinopathy disease level. The ranking of disease level on each candidate set provides a much promising result for developing practically automated and assisted diabetic retinopathy diagnosis system. Experimental work on digital fundus images using the DFIR method performs research on the factors such as sensitivity, ranking efficiency, and feature selection time.

Diagnosing and Ranking Retinopathy Disease Level Using Diabetic Fundus Image Recuperation Approach

PubMed Central

Somasundaram, K.; Alli Rajendran, P.

2015-01-01

Retinal fundus images are widely used in diagnosing different types of eye diseases. The existing methods such as Feature Based Macular Edema Detection (FMED) and Optimally Adjusted Morphological Operator (OAMO) effectively detected the presence of exudation in fundus images and identified the true positive ratio of exudates detection, respectively. These mechanically detected exudates did not include more detailed feature selection technique to the system for detection of diabetic retinopathy. To categorize the exudates, Diabetic Fundus Image Recuperation (DFIR) method based on sliding window approach is developed in this work to select the features of optic cup in digital retinal fundus images. The DFIR feature selection uses collection of sliding windows with varying range to obtain the features based on the histogram value using Group Sparsity Nonoverlapping Function. Using support vector model in the second phase, the DFIR method based on Spiral Basis Function effectively ranks the diabetic retinopathy disease level. The ranking of disease level on each candidate set provides a much promising result for developing practically automated and assisted diabetic retinopathy diagnosis system. Experimental work on digital fundus images using the DFIR method performs research on the factors such as sensitivity, ranking efficiency, and feature selection time. PMID:25945362

Stochastic model search with binary outcomes for genome-wide association studies

PubMed Central

Malovini, Alberto; Puca, Annibale A; Bellazzi, Riccardo

2012-01-01

Objective The spread of case–control genome-wide association studies (GWASs) has stimulated the development of new variable selection methods and predictive models. We introduce a novel Bayesian model search algorithm, Binary Outcome Stochastic Search (BOSS), which addresses the model selection problem when the number of predictors far exceeds the number of binary responses. Materials and methods Our method is based on a latent variable model that links the observed outcomes to the underlying genetic variables. A Markov Chain Monte Carlo approach is used for model search and to evaluate the posterior probability of each predictor. Results BOSS is compared with three established methods (stepwise regression, logistic lasso, and elastic net) in a simulated benchmark. Two real case studies are also investigated: a GWAS on the genetic bases of longevity, and the type 2 diabetes study from the Wellcome Trust Case Control Consortium. Simulations show that BOSS achieves higher precisions than the reference methods while preserving good recall rates. In both experimental studies, BOSS successfully detects genetic polymorphisms previously reported to be associated with the analyzed phenotypes. Discussion BOSS outperforms the other methods in terms of F-measure on simulated data. In the two real studies, BOSS successfully detects biologically relevant features, some of which are missed by univariate analysis and the three reference techniques. Conclusion The proposed algorithm is an advance in the methodology for model selection with a large number of features. Our simulated and experimental results showed that BOSS proves effective in detecting relevant markers while providing a parsimonious model. PMID:22534080

A New Direction of Cancer Classification: Positive Effect of Low-Ranking MicroRNAs.

PubMed

Li, Feifei; Piao, Minghao; Piao, Yongjun; Li, Meijing; Ryu, Keun Ho

2014-10-01

Many studies based on microRNA (miRNA) expression profiles showed a new aspect of cancer classification. Because one characteristic of miRNA expression data is the high dimensionality, feature selection methods have been used to facilitate dimensionality reduction. The feature selection methods have one shortcoming thus far: they just consider the problem of where feature to class is 1:1 or n:1. However, because one miRNA may influence more than one type of cancer, human miRNA is considered to be ranked low in traditional feature selection methods and are removed most of the time. In view of the limitation of the miRNA number, low-ranking miRNAs are also important to cancer classification. We considered both high- and low-ranking features to cover all problems (1:1, n:1, 1:n, and m:n) in cancer classification. First, we used the correlation-based feature selection method to select the high-ranking miRNAs, and chose the support vector machine, Bayes network, decision tree, k-nearest-neighbor, and logistic classifier to construct cancer classification. Then, we chose Chi-square test, information gain, gain ratio, and Pearson's correlation feature selection methods to build the m:n feature subset, and used the selected miRNAs to determine cancer classification. The low-ranking miRNA expression profiles achieved higher classification accuracy compared with just using high-ranking miRNAs in traditional feature selection methods. Our results demonstrate that the m:n feature subset made a positive impression of low-ranking miRNAs in cancer classification.

A robust and efficient stepwise regression method for building sparse polynomial chaos expansions

DOE Office of Scientific and Technical Information (OSTI.GOV)

Abraham, Simon, E-mail: Simon.Abraham@ulb.ac.be; Raisee, Mehrdad; Ghorbaniasl, Ghader

2017-03-01

Polynomial Chaos (PC) expansions are widely used in various engineering fields for quantifying uncertainties arising from uncertain parameters. The computational cost of classical PC solution schemes is unaffordable as the number of deterministic simulations to be calculated grows dramatically with the number of stochastic dimension. This considerably restricts the practical use of PC at the industrial level. A common approach to address such problems is to make use of sparse PC expansions. This paper presents a non-intrusive regression-based method for building sparse PC expansions. The most important PC contributions are detected sequentially through an automatic search procedure. The variable selectionmore » criterion is based on efficient tools relevant to probabilistic method. Two benchmark analytical functions are used to validate the proposed algorithm. The computational efficiency of the method is then illustrated by a more realistic CFD application, consisting of the non-deterministic flow around a transonic airfoil subject to geometrical uncertainties. To assess the performance of the developed methodology, a detailed comparison is made with the well established LAR-based selection technique. The results show that the developed sparse regression technique is able to identify the most significant PC contributions describing the problem. Moreover, the most important stochastic features are captured at a reduced computational cost compared to the LAR method. The results also demonstrate the superior robustness of the method by repeating the analyses using random experimental designs.« less

A Novel Image Recuperation Approach for Diagnosing and Ranking Retinopathy Disease Level Using Diabetic Fundus Image

PubMed Central

2015-01-01

Retinal fundus images are widely used in diagnosing and providing treatment for several eye diseases. Prior works using retinal fundus images detected the presence of exudation with the aid of publicly available dataset using extensive segmentation process. Though it was proved to be computationally efficient, it failed to create a diabetic retinopathy feature selection system for transparently diagnosing the disease state. Also the diagnosis of diseases did not employ machine learning methods to categorize candidate fundus images into true positive and true negative ratio. Several candidate fundus images did not include more detailed feature selection technique for diabetic retinopathy. To apply machine learning methods and classify the candidate fundus images on the basis of sliding window a method called, Diabetic Fundus Image Recuperation (DFIR) is designed in this paper. The initial phase of DFIR method select the feature of optic cup in digital retinal fundus images based on Sliding Window Approach. With this, the disease state for diabetic retinopathy is assessed. The feature selection in DFIR method uses collection of sliding windows to obtain the features based on the histogram value. The histogram based feature selection with the aid of Group Sparsity Non-overlapping function provides more detailed information of features. Using Support Vector Model in the second phase, the DFIR method based on Spiral Basis Function effectively ranks the diabetic retinopathy diseases. The ranking of disease level for each candidate set provides a much promising result for developing practically automated diabetic retinopathy diagnosis system. Experimental work on digital fundus images using the DFIR method performs research on the factors such as sensitivity, specificity rate, ranking efficiency and feature selection time. PMID:25974230

Characterizing multivariate decoding models based on correlated EEG spectral features

PubMed Central

McFarland, Dennis J.

2013-01-01

Objective Multivariate decoding methods are popular techniques for analysis of neurophysiological data. The present study explored potential interpretative problems with these techniques when predictors are correlated. Methods Data from sensorimotor rhythm-based cursor control experiments was analyzed offline with linear univariate and multivariate models. Features were derived from autoregressive (AR) spectral analysis of varying model order which produced predictors that varied in their degree of correlation (i.e., multicollinearity). Results The use of multivariate regression models resulted in much better prediction of target position as compared to univariate regression models. However, with lower order AR features interpretation of the spectral patterns of the weights was difficult. This is likely to be due to the high degree of multicollinearity present with lower order AR features. Conclusions Care should be exercised when interpreting the pattern of weights of multivariate models with correlated predictors. Comparison with univariate statistics is advisable. Significance While multivariate decoding algorithms are very useful for prediction their utility for interpretation may be limited when predictors are correlated. PMID:23466267

A Comparative Investigation of the Combined Effects of Pre-Processing, Wavelength Selection, and Regression Methods on Near-Infrared Calibration Model Performance.

PubMed

Wan, Jian; Chen, Yi-Chieh; Morris, A Julian; Thennadil, Suresh N

2017-07-01

Near-infrared (NIR) spectroscopy is being widely used in various fields ranging from pharmaceutics to the food industry for analyzing chemical and physical properties of the substances concerned. Its advantages over other analytical techniques include available physical interpretation of spectral data, nondestructive nature and high speed of measurements, and little or no need for sample preparation. The successful application of NIR spectroscopy relies on three main aspects: pre-processing of spectral data to eliminate nonlinear variations due to temperature, light scattering effects and many others, selection of those wavelengths that contribute useful information, and identification of suitable calibration models using linear/nonlinear regression . Several methods have been developed for each of these three aspects and many comparative studies of different methods exist for an individual aspect or some combinations. However, there is still a lack of comparative studies for the interactions among these three aspects, which can shed light on what role each aspect plays in the calibration and how to combine various methods of each aspect together to obtain the best calibration model. This paper aims to provide such a comparative study based on four benchmark data sets using three typical pre-processing methods, namely, orthogonal signal correction (OSC), extended multiplicative signal correction (EMSC) and optical path-length estimation and correction (OPLEC); two existing wavelength selection methods, namely, stepwise forward selection (SFS) and genetic algorithm optimization combined with partial least squares regression for spectral data (GAPLSSP); four popular regression methods, namely, partial least squares (PLS), least absolute shrinkage and selection operator (LASSO), least squares support vector machine (LS-SVM), and Gaussian process regression (GPR). The comparative study indicates that, in general, pre-processing of spectral data can play a significant role in the calibration while wavelength selection plays a marginal role and the combination of certain pre-processing, wavelength selection, and nonlinear regression methods can achieve superior performance over traditional linear regression-based calibration.

Learning representations for the early detection of sepsis with deep neural networks.

PubMed

Kam, Hye Jin; Kim, Ha Young

2017-10-01

Sepsis is one of the leading causes of death in intensive care unit patients. Early detection of sepsis is vital because mortality increases as the sepsis stage worsens. This study aimed to develop detection models for the early stage of sepsis using deep learning methodologies, and to compare the feasibility and performance of the new deep learning methodology with those of the regression method with conventional temporal feature extraction. Study group selection adhered to the InSight model. The results of the deep learning-based models and the InSight model were compared. With deep feedforward networks, the area under the ROC curve (AUC) of the models were 0.887 and 0.915 for the InSight and the new feature sets, respectively. For the model with the combined feature set, the AUC was the same as that of the basic feature set (0.915). For the long short-term memory model, only the basic feature set was applied and the AUC improved to 0.929 compared with the existing 0.887 of the InSight model. The contributions of this paper can be summarized in three ways: (i) improved performance without feature extraction using domain knowledge, (ii) verification of feature extraction capability of deep neural networks through comparison with reference features, and (iii) improved performance with feedforward neural networks using long short-term memory, a neural network architecture that can learn sequential patterns. Copyright © 2017 Elsevier Ltd. All rights reserved.

Imaging genetics approach to predict progression of Parkinson's diseases.

PubMed

Mansu Kim; Seong-Jin Son; Hyunjin Park

2017-07-01

Imaging genetics is a tool to extract genetic variants associated with both clinical phenotypes and imaging information. The approach can extract additional genetic variants compared to conventional approaches to better investigate various diseased conditions. Here, we applied imaging genetics to study Parkinson's disease (PD). We aimed to extract significant features derived from imaging genetics and neuroimaging. We built a regression model based on extracted significant features combining genetics and neuroimaging to better predict clinical scores of PD progression (i.e. MDS-UPDRS). Our model yielded high correlation (r = 0.697, p <; 0.001) and low root mean squared error (8.36) between predicted and actual MDS-UPDRS scores. Neuroimaging (from 123 I-Ioflupane SPECT) predictors of regression model were computed from independent component analysis approach. Genetic features were computed using image genetics approach based on identified neuroimaging features as intermediate phenotypes. Joint modeling of neuroimaging and genetics could provide complementary information and thus have the potential to provide further insight into the pathophysiology of PD. Our model included newly found neuroimaging features and genetic variants which need further investigation.

Comparative Performance Analysis of Support Vector Machine, Random Forest, Logistic Regression and k-Nearest Neighbours in Rainbow Trout (Oncorhynchus Mykiss) Classification Using Image-Based Features

PubMed Central

Císař, Petr; Labbé, Laurent; Souček, Pavel; Pelissier, Pablo; Kerneis, Thierry

2018-01-01

The main aim of this study was to develop a new objective method for evaluating the impacts of different diets on the live fish skin using image-based features. In total, one-hundred and sixty rainbow trout (Oncorhynchus mykiss) were fed either a fish-meal based diet (80 fish) or a 100% plant-based diet (80 fish) and photographed using consumer-grade digital camera. Twenty-three colour features and four texture features were extracted. Four different classification methods were used to evaluate fish diets including Random forest (RF), Support vector machine (SVM), Logistic regression (LR) and k-Nearest neighbours (k-NN). The SVM with radial based kernel provided the best classifier with correct classification rate (CCR) of 82% and Kappa coefficient of 0.65. Although the both LR and RF methods were less accurate than SVM, they achieved good classification with CCR 75% and 70% respectively. The k-NN was the least accurate (40%) classification model. Overall, it can be concluded that consumer-grade digital cameras could be employed as the fast, accurate and non-invasive sensor for classifying rainbow trout based on their diets. Furthermore, these was a close association between image-based features and fish diet received during cultivation. These procedures can be used as non-invasive, accurate and precise approaches for monitoring fish status during the cultivation by evaluating diet’s effects on fish skin. PMID:29596375

Comparative Performance Analysis of Support Vector Machine, Random Forest, Logistic Regression and k-Nearest Neighbours in Rainbow Trout (Oncorhynchus Mykiss) Classification Using Image-Based Features.

PubMed

Saberioon, Mohammadmehdi; Císař, Petr; Labbé, Laurent; Souček, Pavel; Pelissier, Pablo; Kerneis, Thierry

2018-03-29

The main aim of this study was to develop a new objective method for evaluating the impacts of different diets on the live fish skin using image-based features. In total, one-hundred and sixty rainbow trout ( Oncorhynchus mykiss ) were fed either a fish-meal based diet (80 fish) or a 100% plant-based diet (80 fish) and photographed using consumer-grade digital camera. Twenty-three colour features and four texture features were extracted. Four different classification methods were used to evaluate fish diets including Random forest (RF), Support vector machine (SVM), Logistic regression (LR) and k -Nearest neighbours ( k -NN). The SVM with radial based kernel provided the best classifier with correct classification rate (CCR) of 82% and Kappa coefficient of 0.65. Although the both LR and RF methods were less accurate than SVM, they achieved good classification with CCR 75% and 70% respectively. The k -NN was the least accurate (40%) classification model. Overall, it can be concluded that consumer-grade digital cameras could be employed as the fast, accurate and non-invasive sensor for classifying rainbow trout based on their diets. Furthermore, these was a close association between image-based features and fish diet received during cultivation. These procedures can be used as non-invasive, accurate and precise approaches for monitoring fish status during the cultivation by evaluating diet's effects on fish skin.

Evaluating the performance of different predictor strategies in regression-based downscaling with a focus on glacierized mountain environments

NASA Astrophysics Data System (ADS)

Hofer, Marlis; Nemec, Johanna

2016-04-01

This study presents first steps towards verifying the hypothesis that uncertainty in global and regional glacier mass simulations can be reduced considerably by reducing the uncertainty in the high-resolution atmospheric input data. To this aim, we systematically explore the potential of different predictor strategies for improving the performance of regression-based downscaling approaches. The investigated local-scale target variables are precipitation, air temperature, wind speed, relative humidity and global radiation, all at a daily time scale. Observations of these target variables are assessed from three sites in geo-environmentally and climatologically very distinct settings, all within highly complex topography and in the close proximity to mountain glaciers: (1) the Vernagtbach station in the Northern European Alps (VERNAGT), (2) the Artesonraju measuring site in the tropical South American Andes (ARTESON), and (3) the Brewster measuring site in the Southern Alps of New Zealand (BREWSTER). As the large-scale predictors, ERA interim reanalysis data are used. In the applied downscaling model training and evaluation procedures, particular emphasis is put on appropriately accounting for the pitfalls of limited and/or patchy observation records that are usually the only (if at all) available data from the glacierized mountain sites. Generalized linear models and beta regression are investigated as alternatives to ordinary least squares regression for the non-Gaussian target variables. By analyzing results for the three different sites, five predictands and for different times of the year, we look for systematic improvements in the downscaling models' skill specifically obtained by (i) using predictor data at the optimum scale rather than the minimum scale of the reanalysis data, (ii) identifying the optimum predictor allocation in the vertical, and (iii) considering multiple (variable, level and/or grid point) predictor options combined with state-of-art empirical feature selection tools. First results show that in particular for air temperature, those downscaling models based on direct predictor selection show comparative skill like those models based on multiple predictors. For all other target variables, however, multiple predictor approaches can considerably outperform those models based on single predictors. Including multiple variable types emerges as the most promising predictor option (in particular for wind speed at all sites), even if the same predictor set is used across the different cases.

Semisupervised Clustering by Iterative Partition and Regression with Neuroscience Applications

PubMed Central

Qian, Guoqi; Wu, Yuehua; Ferrari, Davide; Qiao, Puxue; Hollande, Frédéric

2016-01-01

Regression clustering is a mixture of unsupervised and supervised statistical learning and data mining method which is found in a wide range of applications including artificial intelligence and neuroscience. It performs unsupervised learning when it clusters the data according to their respective unobserved regression hyperplanes. The method also performs supervised learning when it fits regression hyperplanes to the corresponding data clusters. Applying regression clustering in practice requires means of determining the underlying number of clusters in the data, finding the cluster label of each data point, and estimating the regression coefficients of the model. In this paper, we review the estimation and selection issues in regression clustering with regard to the least squares and robust statistical methods. We also provide a model selection based technique to determine the number of regression clusters underlying the data. We further develop a computing procedure for regression clustering estimation and selection. Finally, simulation studies are presented for assessing the procedure, together with analyzing a real data set on RGB cell marking in neuroscience to illustrate and interpret the method. PMID:27212939

Deep ensemble learning of sparse regression models for brain disease diagnosis.

PubMed

Suk, Heung-Il; Lee, Seong-Whan; Shen, Dinggang

2017-04-01

Recent studies on brain imaging analysis witnessed the core roles of machine learning techniques in computer-assisted intervention for brain disease diagnosis. Of various machine-learning techniques, sparse regression models have proved their effectiveness in handling high-dimensional data but with a small number of training samples, especially in medical problems. In the meantime, deep learning methods have been making great successes by outperforming the state-of-the-art performances in various applications. In this paper, we propose a novel framework that combines the two conceptually different methods of sparse regression and deep learning for Alzheimer's disease/mild cognitive impairment diagnosis and prognosis. Specifically, we first train multiple sparse regression models, each of which is trained with different values of a regularization control parameter. Thus, our multiple sparse regression models potentially select different feature subsets from the original feature set; thereby they have different powers to predict the response values, i.e., clinical label and clinical scores in our work. By regarding the response values from our sparse regression models as target-level representations, we then build a deep convolutional neural network for clinical decision making, which thus we call 'Deep Ensemble Sparse Regression Network.' To our best knowledge, this is the first work that combines sparse regression models with deep neural network. In our experiments with the ADNI cohort, we validated the effectiveness of the proposed method by achieving the highest diagnostic accuracies in three classification tasks. We also rigorously analyzed our results and compared with the previous studies on the ADNI cohort in the literature. Copyright © 2017 Elsevier B.V. All rights reserved.

Deep ensemble learning of sparse regression models for brain disease diagnosis

PubMed Central

Suk, Heung-Il; Lee, Seong-Whan; Shen, Dinggang

2018-01-01

Recent studies on brain imaging analysis witnessed the core roles of machine learning techniques in computer-assisted intervention for brain disease diagnosis. Of various machine-learning techniques, sparse regression models have proved their effectiveness in handling high-dimensional data but with a small number of training samples, especially in medical problems. In the meantime, deep learning methods have been making great successes by outperforming the state-of-the-art performances in various applications. In this paper, we propose a novel framework that combines the two conceptually different methods of sparse regression and deep learning for Alzheimer’s disease/mild cognitive impairment diagnosis and prognosis. Specifically, we first train multiple sparse regression models, each of which is trained with different values of a regularization control parameter. Thus, our multiple sparse regression models potentially select different feature subsets from the original feature set; thereby they have different powers to predict the response values, i.e., clinical label and clinical scores in our work. By regarding the response values from our sparse regression models as target-level representations, we then build a deep convolutional neural network for clinical decision making, which thus we call ‘ Deep Ensemble Sparse Regression Network.’ To our best knowledge, this is the first work that combines sparse regression models with deep neural network. In our experiments with the ADNI cohort, we validated the effectiveness of the proposed method by achieving the highest diagnostic accuracies in three classification tasks. We also rigorously analyzed our results and compared with the previous studies on the ADNI cohort in the literature. PMID:28167394

Two Paradoxes in Linear Regression Analysis

PubMed Central

FENG, Ge; PENG, Jing; TU, Dongke; ZHENG, Julia Z.; FENG, Changyong

2016-01-01

Summary Regression is one of the favorite tools in applied statistics. However, misuse and misinterpretation of results from regression analysis are common in biomedical research. In this paper we use statistical theory and simulation studies to clarify some paradoxes around this popular statistical method. In particular, we show that a widely used model selection procedure employed in many publications in top medical journals is wrong. Formal procedures based on solid statistical theory should be used in model selection. PMID:28638214

A universal deep learning approach for modeling the flow of patients under different severities.

PubMed

Jiang, Shancheng; Chin, Kwai-Sang; Tsui, Kwok L

2018-02-01

The Accident and Emergency Department (A&ED) is the frontline for providing emergency care in hospitals. Unfortunately, relative A&ED resources have failed to keep up with continuously increasing demand in recent years, which leads to overcrowding in A&ED. Knowing the fluctuation of patient arrival volume in advance is a significant premise to relieve this pressure. Based on this motivation, the objective of this study is to explore an integrated framework with high accuracy for predicting A&ED patient flow under different triage levels, by combining a novel feature selection process with deep neural networks. Administrative data is collected from an actual A&ED and categorized into five groups based on different triage levels. A genetic algorithm (GA)-based feature selection algorithm is improved and implemented as a pre-processing step for this time-series prediction problem, in order to explore key features affecting patient flow. In our improved GA, a fitness-based crossover is proposed to maintain the joint information of multiple features during iterative process, instead of traditional point-based crossover. Deep neural networks (DNN) is employed as the prediction model to utilize their universal adaptability and high flexibility. In the model-training process, the learning algorithm is well-configured based on a parallel stochastic gradient descent algorithm. Two effective regularization strategies are integrated in one DNN framework to avoid overfitting. All introduced hyper-parameters are optimized efficiently by grid-search in one pass. As for feature selection, our improved GA-based feature selection algorithm has outperformed a typical GA and four state-of-the-art feature selection algorithms (mRMR, SAFS, VIFR, and CFR). As for the prediction accuracy of proposed integrated framework, compared with other frequently used statistical models (GLM, seasonal-ARIMA, ARIMAX, and ANN) and modern machine models (SVM-RBF, SVM-linear, RF, and R-LASSO), the proposed integrated "DNN-I-GA" framework achieves higher prediction accuracy on both MAPE and RMSE metrics in pairwise comparisons. The contribution of our study is two-fold. Theoretically, the traditional GA-based feature selection process is improved to have less hyper-parameters and higher efficiency, and the joint information of multiple features is maintained by fitness-based crossover operator. The universal property of DNN is further enhanced by merging different regularization strategies. Practically, features selected by our improved GA can be used to acquire an underlying relationship between patient flows and input features. Predictive values are significant indicators of patients' demand and can be used by A&ED managers to make resource planning and allocation. High accuracy achieved by the present framework in different cases enhances the reliability of downstream decision makings. Copyright © 2017 Elsevier B.V. All rights reserved.

Tumor recognition in wireless capsule endoscopy images using textural features and SVM-based feature selection.

PubMed

Li, Baopu; Meng, Max Q-H

2012-05-01

Tumor in digestive tract is a common disease and wireless capsule endoscopy (WCE) is a relatively new technology to examine diseases for digestive tract especially for small intestine. This paper addresses the problem of automatic recognition of tumor for WCE images. Candidate color texture feature that integrates uniform local binary pattern and wavelet is proposed to characterize WCE images. The proposed features are invariant to illumination change and describe multiresolution characteristics of WCE images. Two feature selection approaches based on support vector machine, sequential forward floating selection and recursive feature elimination, are further employed to refine the proposed features for improving the detection accuracy. Extensive experiments validate that the proposed computer-aided diagnosis system achieves a promising tumor recognition accuracy of 92.4% in WCE images on our collected data.

Combining the genetic algorithm and successive projection algorithm for the selection of feature wavelengths to evaluate exudative characteristics in frozen-thawed fish muscle.

PubMed

Cheng, Jun-Hu; Sun, Da-Wen; Pu, Hongbin

2016-04-15

The potential use of feature wavelengths for predicting drip loss in grass carp fish, as affected by being frozen at -20°C for 24 h and thawed at 4°C for 1, 2, 4, and 6 days, was investigated. Hyperspectral images of frozen-thawed fish were obtained and their corresponding spectra were extracted. Least-squares support vector machine and multiple linear regression (MLR) models were established using five key wavelengths, selected by combining a genetic algorithm and successive projections algorithm, and this showed satisfactory performance in drip loss prediction. The MLR model with a determination coefficient of prediction (R(2)P) of 0.9258, and lower root mean square error estimated by a prediction (RMSEP) of 1.12%, was applied to transfer each pixel of the image and generate the distribution maps of exudation changes. The results confirmed that it is feasible to identify the feature wavelengths using variable selection methods and chemometric analysis for developing on-line multispectral imaging. Copyright © 2015 Elsevier Ltd. All rights reserved.

Seminal quality prediction using data mining methods.

PubMed

Sahoo, Anoop J; Kumar, Yugal

2014-01-01

Now-a-days, some new classes of diseases have come into existences which are known as lifestyle diseases. The main reasons behind these diseases are changes in the lifestyle of people such as alcohol drinking, smoking, food habits etc. After going through the various lifestyle diseases, it has been found that the fertility rates (sperm quantity) in men has considerably been decreasing in last two decades. Lifestyle factors as well as environmental factors are mainly responsible for the change in the semen quality. The objective of this paper is to identify the lifestyle and environmental features that affects the seminal quality and also fertility rate in man using data mining methods. The five artificial intelligence techniques such as Multilayer perceptron (MLP), Decision Tree (DT), Navie Bayes (Kernel), Support vector machine+Particle swarm optimization (SVM+PSO) and Support vector machine (SVM) have been applied on fertility dataset to evaluate the seminal quality and also to predict the person is either normal or having altered fertility rate. While the eight feature selection techniques such as support vector machine (SVM), neural network (NN), evolutionary logistic regression (LR), support vector machine plus particle swarm optimization (SVM+PSO), principle component analysis (PCA), chi-square test, correlation and T-test methods have been used to identify more relevant features which affect the seminal quality. These techniques are applied on fertility dataset which contains 100 instances with nine attribute with two classes. The experimental result shows that SVM+PSO provides higher accuracy and area under curve (AUC) rate (94% & 0.932) among multi-layer perceptron (MLP) (92% & 0.728), Support Vector Machines (91% & 0.758), Navie Bayes (Kernel) (89% & 0.850) and Decision Tree (89% & 0.735) for some of the seminal parameters. This paper also focuses on the feature selection process i.e. how to select the features which are more important for prediction of fertility rate. In this paper, eight feature selection methods are applied on fertility dataset to find out a set of good features. The investigational results shows that childish diseases (0.079) and high fever features (0.057) has less impact on fertility rate while age (0.8685), season (0.843), surgical intervention (0.7683), alcohol consumption (0.5992), smoking habit (0.575), number of hours spent on setting (0.4366) and accident (0.5973) features have more impact. It is also observed that feature selection methods increase the accuracy of above mentioned techniques (multilayer perceptron 92%, support vector machine 91%, SVM+PSO 94%, Navie Bayes (Kernel) 89% and decision tree 89%) as compared to without feature selection methods (multilayer perceptron 86%, support vector machine 86%, SVM+PSO 85%, Navie Bayes (Kernel) 83% and decision tree 84%) which shows the applicability of feature selection methods in prediction. This paper lightens the application of artificial techniques in medical domain. From this paper, it can be concluded that data mining methods can be used to predict a person with or without disease based on environmental and lifestyle parameters/features rather than undergoing various medical test. In this paper, five data mining techniques are used to predict the fertility rate and among which SVM+PSO provide more accurate results than support vector machine and decision tree.

QSRR modeling for the chromatographic retention behavior of some β-lactam antibiotics using forward and firefly variable selection algorithms coupled with multiple linear regression.

PubMed

Fouad, Marwa A; Tolba, Enas H; El-Shal, Manal A; El Kerdawy, Ahmed M

2018-05-11

The justified continuous emerging of new β-lactam antibiotics provokes the need for developing suitable analytical methods that accelerate and facilitate their analysis. A face central composite experimental design was adopted using different levels of phosphate buffer pH, acetonitrile percentage at zero time and after 15 min in a gradient program to obtain the optimum chromatographic conditions for the elution of 31 β-lactam antibiotics. Retention factors were used as the target property to build two QSRR models utilizing the conventional forward selection and the advanced nature-inspired firefly algorithm for descriptor selection, coupled with multiple linear regression. The obtained models showed high performance in both internal and external validation indicating their robustness and predictive ability. Williams-Hotelling test and student's t-test showed that there is no statistical significant difference between the models' results. Y-randomization validation showed that the obtained models are due to significant correlation between the selected molecular descriptors and the analytes' chromatographic retention. These results indicate that the generated FS-MLR and FFA-MLR models are showing comparable quality on both the training and validation levels. They also gave comparable information about the molecular features that influence the retention behavior of β-lactams under the current chromatographic conditions. We can conclude that in some cases simple conventional feature selection algorithm can be used to generate robust and predictive models comparable to that are generated using advanced ones. Copyright © 2018 Elsevier B.V. All rights reserved.

Optimization of breast mass classification using sequential forward floating selection (SFFS) and a support vector machine (SVM) model

PubMed Central

Tan, Maxine; Pu, Jiantao; Zheng, Bin

2014-01-01

Purpose: Improving radiologists’ performance in classification between malignant and benign breast lesions is important to increase cancer detection sensitivity and reduce false-positive recalls. For this purpose, developing computer-aided diagnosis (CAD) schemes has been attracting research interest in recent years. In this study, we investigated a new feature selection method for the task of breast mass classification. Methods: We initially computed 181 image features based on mass shape, spiculation, contrast, presence of fat or calcifications, texture, isodensity, and other morphological features. From this large image feature pool, we used a sequential forward floating selection (SFFS)-based feature selection method to select relevant features, and analyzed their performance using a support vector machine (SVM) model trained for the classification task. On a database of 600 benign and 600 malignant mass regions of interest (ROIs), we performed the study using a ten-fold cross-validation method. Feature selection and optimization of the SVM parameters were conducted on the training subsets only. Results: The area under the receiver operating characteristic curve (AUC) = 0.805±0.012 was obtained for the classification task. The results also showed that the most frequently-selected features by the SFFS-based algorithm in 10-fold iterations were those related to mass shape, isodensity and presence of fat, which are consistent with the image features frequently used by radiologists in the clinical environment for mass classification. The study also indicated that accurately computing mass spiculation features from the projection mammograms was difficult, and failed to perform well for the mass classification task due to tissue overlap within the benign mass regions. Conclusions: In conclusion, this comprehensive feature analysis study provided new and valuable information for optimizing computerized mass classification schemes that may have potential to be useful as a “second reader” in future clinical practice. PMID:24664267

The optional selection of micro-motion feature based on Support Vector Machine

NASA Astrophysics Data System (ADS)

Li, Bo; Ren, Hongmei; Xiao, Zhi-he; Sheng, Jing

2017-11-01

Micro-motion form of target is multiple, different micro-motion forms are apt to be modulated, which makes it difficult for feature extraction and recognition. Aiming at feature extraction of cone-shaped objects with different micro-motion forms, this paper proposes the best selection method of micro-motion feature based on support vector machine. After the time-frequency distribution of radar echoes, comparing the time-frequency spectrum of objects with different micro-motion forms, features are extracted based on the differences between the instantaneous frequency variations of different micro-motions. According to the methods based on SVM (Support Vector Machine) features are extracted, then the best features are acquired. Finally, the result shows the method proposed in this paper is feasible under the test condition of certain signal-to-noise ratio(SNR).

A ROC-based feature selection method for computer-aided detection and diagnosis

NASA Astrophysics Data System (ADS)

Wang, Songyuan; Zhang, Guopeng; Liao, Qimei; Zhang, Junying; Jiao, Chun; Lu, Hongbing

2014-03-01

Image-based computer-aided detection and diagnosis (CAD) has been a very active research topic aiming to assist physicians to detect lesions and distinguish them from benign to malignant. However, the datasets fed into a classifier usually suffer from small number of samples, as well as significantly less samples available in one class (have a disease) than the other, resulting in the classifier's suboptimal performance. How to identifying the most characterizing features of the observed data for lesion detection is critical to improve the sensitivity and minimize false positives of a CAD system. In this study, we propose a novel feature selection method mR-FAST that combines the minimal-redundancymaximal relevance (mRMR) framework with a selection metric FAST (feature assessment by sliding thresholds) based on the area under a ROC curve (AUC) generated on optimal simple linear discriminants. With three feature datasets extracted from CAD systems for colon polyps and bladder cancer, we show that the space of candidate features selected by mR-FAST is more characterizing for lesion detection with higher AUC, enabling to find a compact subset of superior features at low cost.

Predicting Retention Times of Naturally Occurring Phenolic Compounds in Reversed-Phase Liquid Chromatography: A Quantitative Structure-Retention Relationship (QSRR) Approach

PubMed Central

Akbar, Jamshed; Iqbal, Shahid; Batool, Fozia; Karim, Abdul; Chan, Kim Wei

2012-01-01

Quantitative structure-retention relationships (QSRRs) have successfully been developed for naturally occurring phenolic compounds in a reversed-phase liquid chromatographic (RPLC) system. A total of 1519 descriptors were calculated from the optimized structures of the molecules using MOPAC2009 and DRAGON softwares. The data set of 39 molecules was divided into training and external validation sets. For feature selection and mapping we used step-wise multiple linear regression (SMLR), unsupervised forward selection followed by step-wise multiple linear regression (UFS-SMLR) and artificial neural networks (ANN). Stable and robust models with significant predictive abilities in terms of validation statistics were obtained with negation of any chance correlation. ANN models were found better than remaining two approaches. HNar, IDM, Mp, GATS2v, DISP and 3D-MoRSE (signals 22, 28 and 32) descriptors based on van der Waals volume, electronegativity, mass and polarizability, at atomic level, were found to have significant effects on the retention times. The possible implications of these descriptors in RPLC have been discussed. All the models are proven to be quite able to predict the retention times of phenolic compounds and have shown remarkable validation, robustness, stability and predictive performance. PMID:23203132

Conversion Discriminative Analysis on Mild Cognitive Impairment Using Multiple Cortical Features from MR Images.

PubMed

Guo, Shengwen; Lai, Chunren; Wu, Congling; Cen, Guiyin

2017-01-01

Neuroimaging measurements derived from magnetic resonance imaging provide important information required for detecting changes related to the progression of mild cognitive impairment (MCI). Cortical features and changes play a crucial role in revealing unique anatomical patterns of brain regions, and further differentiate MCI patients from normal states. Four cortical features, namely, gray matter volume, cortical thickness, surface area, and mean curvature, were explored for discriminative analysis among three groups including the stable MCI (sMCI), the converted MCI (cMCI), and the normal control (NC) groups. In this study, 158 subjects (72 NC, 46 sMCI, and 40 cMCI) were selected from the Alzheimer's Disease Neuroimaging Initiative. A sparse-constrained regression model based on the l2-1-norm was introduced to reduce the feature dimensionality and retrieve essential features for the discrimination of the three groups by using a support vector machine (SVM). An optimized strategy of feature addition based on the weight of each feature was adopted for the SVM classifier in order to achieve the best classification performance. The baseline cortical features combined with the longitudinal measurements for 2 years of follow-up data yielded prominent classification results. In particular, the cortical thickness produced a classification with 98.84% accuracy, 97.5% sensitivity, and 100% specificity for the sMCI-cMCI comparison; 92.37% accuracy, 84.78% sensitivity, and 97.22% specificity for the cMCI-NC comparison; and 93.75% accuracy, 92.5% sensitivity, and 94.44% specificity for the sMCI-NC comparison. The best performances obtained by the SVM classifier using the essential features were 5-40% more than those using all of the retained features. The feasibility of the cortical features for the recognition of anatomical patterns was certified; thus, the proposed method has the potential to improve the clinical diagnosis of sub-types of MCI and predict the risk of its conversion to Alzheimer's disease.

Robust mislabel logistic regression without modeling mislabel probabilities.

PubMed

Hung, Hung; Jou, Zhi-Yu; Huang, Su-Yun

2018-03-01

Logistic regression is among the most widely used statistical methods for linear discriminant analysis. In many applications, we only observe possibly mislabeled responses. Fitting a conventional logistic regression can then lead to biased estimation. One common resolution is to fit a mislabel logistic regression model, which takes into consideration of mislabeled responses. Another common method is to adopt a robust M-estimation by down-weighting suspected instances. In this work, we propose a new robust mislabel logistic regression based on γ-divergence. Our proposal possesses two advantageous features: (1) It does not need to model the mislabel probabilities. (2) The minimum γ-divergence estimation leads to a weighted estimating equation without the need to include any bias correction term, that is, it is automatically bias-corrected. These features make the proposed γ-logistic regression more robust in model fitting and more intuitive for model interpretation through a simple weighting scheme. Our method is also easy to implement, and two types of algorithms are included. Simulation studies and the Pima data application are presented to demonstrate the performance of γ-logistic regression. © 2017, The International Biometric Society.

Prediction of protein-protein interactions based on PseAA composition and hybrid feature selection.

PubMed

Liu, Liang; Cai, Yudong; Lu, Wencong; Feng, Kaiyan; Peng, Chunrong; Niu, Bing

2009-03-06

Based on pseudo amino acid (PseAA) composition and a novel hybrid feature selection frame, this paper presents a computational system to predict the PPIs (protein-protein interactions) using 8796 protein pairs. These pairs are coded by PseAA composition, resulting in 114 features. A hybrid feature selection system, mRMR-KNNs-wrapper, is applied to obtain an optimized feature set by excluding poor-performed and/or redundant features, resulting in 103 remaining features. Using the optimized 103-feature subset, a prediction model is trained and tested in the k-nearest neighbors (KNNs) learning system. This prediction model achieves an overall accurate prediction rate of 76.18%, evaluated by 10-fold cross-validation test, which is 1.46% higher than using the initial 114 features and is 6.51% higher than the 20 features, coded by amino acid compositions. The PPIs predictor, developed for this research, is available for public use at http://chemdata.shu.edu.cn/ppi.

Automatic parameter selection for feature-based multi-sensor image registration

NASA Astrophysics Data System (ADS)

DelMarco, Stephen; Tom, Victor; Webb, Helen; Chao, Alan

2006-05-01

Accurate image registration is critical for applications such as precision targeting, geo-location, change-detection, surveillance, and remote sensing. However, the increasing volume of image data is exceeding the current capacity of human analysts to perform manual registration. This image data glut necessitates the development of automated approaches to image registration, including algorithm parameter value selection. Proper parameter value selection is crucial to the success of registration techniques. The appropriate algorithm parameters can be highly scene and sensor dependent. Therefore, robust algorithm parameter value selection approaches are a critical component of an end-to-end image registration algorithm. In previous work, we developed a general framework for multisensor image registration which includes feature-based registration approaches. In this work we examine the problem of automated parameter selection. We apply the automated parameter selection approach of Yitzhaky and Peli to select parameters for feature-based registration of multisensor image data. The approach consists of generating multiple feature-detected images by sweeping over parameter combinations and using these images to generate estimated ground truth. The feature-detected images are compared to the estimated ground truth images to generate ROC points associated with each parameter combination. We develop a strategy for selecting the optimal parameter set by choosing the parameter combination corresponding to the optimal ROC point. We present numerical results showing the effectiveness of the approach using registration of collected SAR data to reference EO data.

Comparison of Object-Based Image Analysis Approaches to Mapping New Buildings in Accra, Ghana Using Multi-Temporal QuickBird Satellite Imagery

PubMed Central

Tsai, Yu Hsin; Stow, Douglas; Weeks, John

2013-01-01

The goal of this study was to map and quantify the number of newly constructed buildings in Accra, Ghana between 2002 and 2010 based on high spatial resolution satellite image data. Two semi-automated feature detection approaches for detecting and mapping newly constructed buildings based on QuickBird very high spatial resolution satellite imagery were analyzed: (1) post-classification comparison; and (2) bi-temporal layerstack classification. Feature Analyst software based on a spatial contextual classifier and ENVI Feature Extraction that uses a true object-based image analysis approach of image segmentation and segment classification were evaluated. Final map products representing new building objects were compared and assessed for accuracy using two object-based accuracy measures, completeness and correctness. The bi-temporal layerstack method generated more accurate results compared to the post-classification comparison method due to less confusion with background objects. The spectral/spatial contextual approach (Feature Analyst) outperformed the true object-based feature delineation approach (ENVI Feature Extraction) due to its ability to more reliably delineate individual buildings of various sizes. Semi-automated, object-based detection followed by manual editing appears to be a reliable and efficient approach for detecting and enumerating new building objects. A bivariate regression analysis was performed using neighborhood-level estimates of new building density regressed on a census-derived measure of socio-economic status, yielding an inverse relationship with R2 = 0.31 (n = 27; p = 0.00). The primary utility of the new building delineation results is to support spatial analyses of land cover and land use and demographic change. PMID:24415810

Analysis of training sample selection strategies for regression-based quantitative landslide susceptibility mapping methods

NASA Astrophysics Data System (ADS)

Erener, Arzu; Sivas, A. Abdullah; Selcuk-Kestel, A. Sevtap; Düzgün, H. Sebnem

2017-07-01

All of the quantitative landslide susceptibility mapping (QLSM) methods requires two basic data types, namely, landslide inventory and factors that influence landslide occurrence (landslide influencing factors, LIF). Depending on type of landslides, nature of triggers and LIF, accuracy of the QLSM methods differs. Moreover, how to balance the number of 0 (nonoccurrence) and 1 (occurrence) in the training set obtained from the landslide inventory and how to select which one of the 1's and 0's to be included in QLSM models play critical role in the accuracy of the QLSM. Although performance of various QLSM methods is largely investigated in the literature, the challenge of training set construction is not adequately investigated for the QLSM methods. In order to tackle this challenge, in this study three different training set selection strategies along with the original data set is used for testing the performance of three different regression methods namely Logistic Regression (LR), Bayesian Logistic Regression (BLR) and Fuzzy Logistic Regression (FLR). The first sampling strategy is proportional random sampling (PRS), which takes into account a weighted selection of landslide occurrences in the sample set. The second method, namely non-selective nearby sampling (NNS), includes randomly selected sites and their surrounding neighboring points at certain preselected distances to include the impact of clustering. Selective nearby sampling (SNS) is the third method, which concentrates on the group of 1's and their surrounding neighborhood. A randomly selected group of landslide sites and their neighborhood are considered in the analyses similar to NNS parameters. It is found that LR-PRS, FLR-PRS and BLR-Whole Data set-ups, with order, yield the best fits among the other alternatives. The results indicate that in QLSM based on regression models, avoidance of spatial correlation in the data set is critical for the model's performance.

A modified artificial neural network based prediction technique for tropospheric radio refractivity

PubMed Central

Javeed, Shumaila; Javed, Wajahat; Atif, M.; Uddin, Mueen

2018-01-01

Radio refractivity plays a significant role in the development and design of radio systems for attaining the best level of performance. Refractivity in the troposphere is one of the features affecting electromagnetic waves, and hence the communication system interrupts. In this work, a modified artificial neural network (ANN) based model is applied to predict the refractivity. The suggested ANN model comprises three modules: the data preparation module, the feature selection module, and the forecast module. The first module applies pre-processing to make the data compatible for the feature selection module. The second module discards irrelevant and redundant data from the input set. The third module uses ANN for prediction. The ANN model applies a sigmoid activation function and a multi-variate auto regressive model to update the weights during the training process. In this work, the refractivity is predicted and estimated based on ten years (2002–2011) of meteorological data, such as the temperature, pressure, and humidity, obtained from the Pakistan Meteorological Department (PMD), Islamabad. The refractivity is estimated using the method suggested by the International Telecommunication Union (ITU). The refractivity is predicted for the year 2012 using the database of the previous ten years, with the help of ANN. The ANN model is implemented in MATLAB. Next, the estimated and predicted refractivity levels are validated against each other. The predicted and actual values (PMD data) of the atmospheric parameters agree with each other well, and demonstrate the accuracy of the proposed ANN method. It was further found that all parameters have a strong relationship with refractivity, in particular the temperature and humidity. The refractivity values are higher during the rainy season owing to a strong association with the relative humidity. Therefore, it is important to properly cater the signal communication system during hot and humid weather. Based on the results, the proposed ANN method can be used to develop a refractivity database, which is highly important in a radio communication system. PMID:29494609

Vibration and acoustic frequency spectra for industrial process modeling using selective fusion multi-condition samples and multi-source features

NASA Astrophysics Data System (ADS)

Tang, Jian; Qiao, Junfei; Wu, ZhiWei; Chai, Tianyou; Zhang, Jian; Yu, Wen

2018-01-01

Frequency spectral data of mechanical vibration and acoustic signals relate to difficult-to-measure production quality and quantity parameters of complex industrial processes. A selective ensemble (SEN) algorithm can be used to build a soft sensor model of these process parameters by fusing valued information selectively from different perspectives. However, a combination of several optimized ensemble sub-models with SEN cannot guarantee the best prediction model. In this study, we use several techniques to construct mechanical vibration and acoustic frequency spectra of a data-driven industrial process parameter model based on selective fusion multi-condition samples and multi-source features. Multi-layer SEN (MLSEN) strategy is used to simulate the domain expert cognitive process. Genetic algorithm and kernel partial least squares are used to construct the inside-layer SEN sub-model based on each mechanical vibration and acoustic frequency spectral feature subset. Branch-and-bound and adaptive weighted fusion algorithms are integrated to select and combine outputs of the inside-layer SEN sub-models. Then, the outside-layer SEN is constructed. Thus, "sub-sampling training examples"-based and "manipulating input features"-based ensemble construction methods are integrated, thereby realizing the selective information fusion process based on multi-condition history samples and multi-source input features. This novel approach is applied to a laboratory-scale ball mill grinding process. A comparison with other methods indicates that the proposed MLSEN approach effectively models mechanical vibration and acoustic signals.

Novel 3D ultrasound image-based biomarkers based on a feature selection from a 2D standardized vessel wall thickness map: a tool for sensitive assessment of therapies for carotid atherosclerosis

NASA Astrophysics Data System (ADS)

Chiu, Bernard; Li, Bing; Chow, Tommy W. S.

2013-09-01

With the advent of new therapies and management strategies for carotid atherosclerosis, there is a parallel need for measurement tools or biomarkers to evaluate the efficacy of these new strategies. 3D ultrasound has been shown to provide reproducible measurements of plaque area/volume and vessel wall volume. However, since carotid atherosclerosis is a focal disease that predominantly occurs at bifurcations, biomarkers based on local plaque change may be more sensitive than global volumetric measurements in demonstrating efficacy of new therapies. The ultimate goal of this paper is to develop a biomarker that is based on the local distribution of vessel-wall-plus-plaque thickness change (VWT-Change) that has occurred during the course of a clinical study. To allow comparison between different treatment groups, the VWT-Change distribution of each subject must first be mapped to a standardized domain. In this study, we developed a technique to map the 3D VWT-Change distribution to a 2D standardized template. We then applied a feature selection technique to identify regions on the 2D standardized map on which subjects in different treatment groups exhibit greater difference in VWT-Change. The proposed algorithm was applied to analyse the VWT-Change of 20 subjects in a placebo-controlled study of the effect of atorvastatin (Lipitor). The average VWT-Change for each subject was computed (i) over all points in the 2D map and (ii) over feature points only. For the average computed over all points, 97 subjects per group would be required to detect an effect size of 25% that of atorvastatin in a six-month study. The sample size is reduced to 25 subjects if the average were computed over feature points only. The introduction of this sensitive quantification technique for carotid atherosclerosis progression/regression would allow many proof-of-principle studies to be performed before a more costly and longer study involving a larger population is held to confirm the treatment efficacy.

NetProt: Complex-based Feature Selection.

PubMed

Goh, Wilson Wen Bin; Wong, Limsoon

2017-08-04

Protein complex-based feature selection (PCBFS) provides unparalleled reproducibility with high phenotypic relevance on proteomics data. Currently, there are five PCBFS paradigms, but not all representative methods have been implemented or made readily available. To allow general users to take advantage of these methods, we developed the R-package NetProt, which provides implementations of representative feature-selection methods. NetProt also provides methods for generating simulated differential data and generating pseudocomplexes for complex-based performance benchmarking. The NetProt open source R package is available for download from https://github.com/gohwils/NetProt/releases/ , and online documentation is available at http://rpubs.com/gohwils/204259 .

Assessment of global and local region-based bilateral mammographic feature asymmetry to predict short-term breast cancer risk

NASA Astrophysics Data System (ADS)

Li, Yane; Fan, Ming; Cheng, Hu; Zhang, Peng; Zheng, Bin; Li, Lihua

2018-01-01

This study aims to develop and test a new imaging marker-based short-term breast cancer risk prediction model. An age-matched dataset of 566 screening mammography cases was used. All ‘prior’ images acquired in the two screening series were negative, while in the ‘current’ screening images, 283 cases were positive for cancer and 283 cases remained negative. For each case, two bilateral cranio-caudal view mammograms acquired from the ‘prior’ negative screenings were selected and processed by a computer-aided image processing scheme, which segmented the entire breast area into nine strip-based local regions, extracted the element regions using difference of Gaussian filters, and computed both global- and local-based bilateral asymmetrical image features. An initial feature pool included 190 features related to the spatial distribution and structural similarity of grayscale values, as well as of the magnitude and phase responses of multidirectional Gabor filters. Next, a short-term breast cancer risk prediction model based on a generalized linear model was built using an embedded stepwise regression analysis method to select features and a leave-one-case-out cross-validation method to predict the likelihood of each woman having image-detectable cancer in the next sequential mammography screening. The area under the receiver operating characteristic curve (AUC) values significantly increased from 0.5863  ±  0.0237 to 0.6870  ±  0.0220 when the model trained by the image features extracted from the global regions and by the features extracted from both the global and the matched local regions (p  =  0.0001). The odds ratio values monotonically increased from 1.00-8.11 with a significantly increasing trend in slope (p  =  0.0028) as the model-generated risk score increased. In addition, the AUC values were 0.6555  ±  0.0437, 0.6958  ±  0.0290, and 0.7054  ±  0.0529 for the three age groups of 37-49, 50-65, and 66-87 years old, respectively. AUC values of 0.6529  ±  0.1100, 0.6820  ±  0.0353, 0.6836  ±  0.0302 and 0.8043  ±  0.1067 were yielded for the four mammography density sub-groups (BIRADS from 1-4), respectively. This study demonstrated that bilateral asymmetry features extracted from local regions combined with the global region in bilateral negative mammograms could be used as a new imaging marker to assist in the prediction of short-term breast cancer risk.

Regression equations for estimating concentrations of selected water-quality constituents for selected gaging stations in the Red River of the North Basin, North Dakota, Minnesota, and South Dakota

USGS Publications Warehouse

Williams-Sether, Tara

2004-01-01

The Dakota Water Resources Act, passed by the U.S. Congress on December 15, 2000, authorized the Secretary of the Interior to conduct a comprehensive study of future water-quantity and quality needs of the Red River of the North Basin in North Dakota and possible options to meet those water needs. Previous Red River of the North Basin studies conducted by the Bureau of Reclamation used streamflow and water-quality data bases developed by the U.S. Geological Survey that included data for 1931-84. As a result of the recent congressional authorization and results of previous studies by the Bureau of Reclamation, redevelopment of the streamflow and water-quality data bases with current data through 1999 are needed in order to evaluate and predict the water-quantity and quality effects within the Red River of the North Basin. This report provides updated statistical summaries of selected water-quality constituents and streamflow and the regression relations between them.  Available data for 1931-99 were used to develop regression equations between 5 selected water-quality constituents and streamflow for 38 gaging stations in the Red River of the North Basin. The water-quality constituents that were regressed against streamflow were hardness (as CaCO3), sodium, chloride, sulfate, and dissolved solids. Statistical summaries of the selected water-quality constituents and streamflow for the gaging stations used in the regression equations development and the applications and limitations of the regression equations are presented in this report.

Moving object detection and tracking in videos through turbulent medium

NASA Astrophysics Data System (ADS)

Halder, Kalyan Kumar; Tahtali, Murat; Anavatti, Sreenatha G.

2016-06-01

This paper addresses the problem of identifying and tracking moving objects in a video sequence having a time-varying background. This is a fundamental task in many computer vision applications, though a very challenging one because of turbulence that causes blurring and spatiotemporal movements of the background images. Our proposed approach involves two major steps. First, a moving object detection algorithm that deals with the detection of real motions by separating the turbulence-induced motions using a two-level thresholding technique is used. In the second step, a feature-based generalized regression neural network is applied to track the detected objects throughout the frames in the video sequence. The proposed approach uses the centroid and area features of the moving objects and creates the reference regions instantly by selecting the objects within a circle. Simulation experiments are carried out on several turbulence-degraded video sequences and comparisons with an earlier method confirms that the proposed approach provides a more effective tracking of the targets.

Non-destructive assessment of instrumental and sensory tenderness of lamb meat using NIR hyperspectral imaging.

PubMed

Kamruzzaman, Mohammed; Elmasry, Gamal; Sun, Da-Wen; Allen, Paul

2013-11-01

The purpose of this study was to develop and test a hyperspectral imaging system (900-1700 nm) to predict instrumental and sensory tenderness of lamb meat. Warner-Bratzler shear force (WBSF) values and sensory scores by trained panellists were collected as the indicator of instrumental and sensory tenderness, respectively. Partial least squares regression models were developed for predicting instrumental and sensory tenderness with reasonable accuracy (Rcv=0.84 for WBSF and 0.69 for sensory tenderness). Overall, the results confirmed that the spectral data could become an interesting screening tool to quickly categorise lamb steaks in good (i.e. tender) and bad (i.e. tough) based on WBSF values and sensory scores with overall accuracy of about 94.51% and 91%, respectively. Successive projections algorithm (SPA) was used to select the most important wavelengths for WBSF prediction. Additionally, textural features from Gray Level Co-occurrence Matrix (GLCM) were extracted to determine the correlation between textural features and WBSF values. Copyright © 2013 Elsevier Ltd. All rights reserved.

A comparative study between nonlinear regression and artificial neural network approaches for modelling wild oat (Avena fatua) field emergence

USDA-ARS?s Scientific Manuscript database

Non-linear regression techniques are used widely to fit weed field emergence patterns to soil microclimatic indices using S-type functions. Artificial neural networks present interesting and alternative features for such modeling purposes. In this work, a univariate hydrothermal-time based Weibull m...

Order Selection for General Expression of Nonlinear Autoregressive Model Based on Multivariate Stepwise Regression

NASA Astrophysics Data System (ADS)

Shi, Jinfei; Zhu, Songqing; Chen, Ruwen

2017-12-01

An order selection method based on multiple stepwise regressions is proposed for General Expression of Nonlinear Autoregressive model which converts the model order problem into the variable selection of multiple linear regression equation. The partial autocorrelation function is adopted to define the linear term in GNAR model. The result is set as the initial model, and then the nonlinear terms are introduced gradually. Statistics are chosen to study the improvements of both the new introduced and originally existed variables for the model characteristics, which are adopted to determine the model variables to retain or eliminate. So the optimal model is obtained through data fitting effect measurement or significance test. The simulation and classic time-series data experiment results show that the method proposed is simple, reliable and can be applied to practical engineering.

Dynamic Network Logistic Regression: A Logistic Choice Analysis of Inter- and Intra-Group Blog Citation Dynamics in the 2004 US Presidential Election

PubMed Central

2013-01-01

Methods for analysis of network dynamics have seen great progress in the past decade. This article shows how Dynamic Network Logistic Regression techniques (a special case of the Temporal Exponential Random Graph Models) can be used to implement decision theoretic models for network dynamics in a panel data context. We also provide practical heuristics for model building and assessment. We illustrate the power of these techniques by applying them to a dynamic blog network sampled during the 2004 US presidential election cycle. This is a particularly interesting case because it marks the debut of Internet-based media such as blogs and social networking web sites as institutionally recognized features of the American political landscape. Using a longitudinal sample of all Democratic National Convention/Republican National Convention–designated blog citation networks, we are able to test the influence of various strategic, institutional, and balance-theoretic mechanisms as well as exogenous factors such as seasonality and political events on the propensity of blogs to cite one another over time. Using a combination of deviance-based model selection criteria and simulation-based model adequacy tests, we identify the combination of processes that best characterizes the choice behavior of the contending blogs. PMID:24143060

Beef quality grading using machine vision

NASA Astrophysics Data System (ADS)

Jeyamkondan, S.; Ray, N.; Kranzler, Glenn A.; Biju, Nisha

2000-12-01

A video image analysis system was developed to support automation of beef quality grading. Forty images of ribeye steaks were acquired. Fat and lean meat were differentiated using a fuzzy c-means clustering algorithm. Muscle longissimus dorsi (l.d.) was segmented from the ribeye using morphological operations. At the end of each iteration of erosion and dilation, a convex hull was fitted to the image and compactness was measured. The number of iterations was selected to yield the most compact l.d. Match between the l.d. muscle traced by an expert grader and that segmented by the program was 95.9%. Marbling and color features were extracted from the l.d. muscle and were used to build regression models to predict marbling and color scores. Quality grade was predicted using another regression model incorporating all features. Grades predicted by the model were statistically equivalent to the grades assigned by expert graders.

Feature-based attentional modulations in the absence of direct visual stimulation.

PubMed

Serences, John T; Boynton, Geoffrey M

2007-07-19

When faced with a crowded visual scene, observers must selectively attend to behaviorally relevant objects to avoid sensory overload. Often this selection process is guided by prior knowledge of a target-defining feature (e.g., the color red when looking for an apple), which enhances the firing rate of visual neurons that are selective for the attended feature. Here, we used functional magnetic resonance imaging and a pattern classification algorithm to predict the attentional state of human observers as they monitored a visual feature (one of two directions of motion). We find that feature-specific attention effects spread across the visual field-even to regions of the scene that do not contain a stimulus. This spread of feature-based attention to empty regions of space may facilitate the perception of behaviorally relevant stimuli by increasing sensitivity to attended features at all locations in the visual field.

Optimum location of external markers using feature selection algorithms for real‐time tumor tracking in external‐beam radiotherapy: a virtual phantom study

PubMed Central

Nankali, Saber; Miandoab, Payam Samadi; Baghizadeh, Amin

2016-01-01

In external‐beam radiotherapy, using external markers is one of the most reliable tools to predict tumor position, in clinical applications. The main challenge in this approach is tumor motion tracking with highest accuracy that depends heavily on external markers location, and this issue is the objective of this study. Four commercially available feature selection algorithms entitled 1) Correlation‐based Feature Selection, 2) Classifier, 3) Principal Components, and 4) Relief were proposed to find optimum location of external markers in combination with two “Genetic” and “Ranker” searching procedures. The performance of these algorithms has been evaluated using four‐dimensional extended cardiac‐torso anthropomorphic phantom. Six tumors in lung, three tumors in liver, and 49 points on the thorax surface were taken into account to simulate internal and external motions, respectively. The root mean square error of an adaptive neuro‐fuzzy inference system (ANFIS) as prediction model was considered as metric for quantitatively evaluating the performance of proposed feature selection algorithms. To do this, the thorax surface region was divided into nine smaller segments and predefined tumors motion was predicted by ANFIS using external motion data of given markers at each small segment, separately. Our comparative results showed that all feature selection algorithms can reasonably select specific external markers from those segments where the root mean square error of the ANFIS model is minimum. Moreover, the performance accuracy of proposed feature selection algorithms was compared, separately. For this, each tumor motion was predicted using motion data of those external markers selected by each feature selection algorithm. Duncan statistical test, followed by F‐test, on final results reflected that all proposed feature selection algorithms have the same performance accuracy for lung tumors. But for liver tumors, a correlation‐based feature selection algorithm, in combination with a genetic search algorithm, proved to yield best performance accuracy for selecting optimum markers. PACS numbers: 87.55.km, 87.56.Fc PMID:26894358

Optimum location of external markers using feature selection algorithms for real-time tumor tracking in external-beam radiotherapy: a virtual phantom study.

PubMed

Nankali, Saber; Torshabi, Ahmad Esmaili; Miandoab, Payam Samadi; Baghizadeh, Amin

2016-01-08

In external-beam radiotherapy, using external markers is one of the most reliable tools to predict tumor position, in clinical applications. The main challenge in this approach is tumor motion tracking with highest accuracy that depends heavily on external markers location, and this issue is the objective of this study. Four commercially available feature selection algorithms entitled 1) Correlation-based Feature Selection, 2) Classifier, 3) Principal Components, and 4) Relief were proposed to find optimum location of external markers in combination with two "Genetic" and "Ranker" searching procedures. The performance of these algorithms has been evaluated using four-dimensional extended cardiac-torso anthropomorphic phantom. Six tumors in lung, three tumors in liver, and 49 points on the thorax surface were taken into account to simulate internal and external motions, respectively. The root mean square error of an adaptive neuro-fuzzy inference system (ANFIS) as prediction model was considered as metric for quantitatively evaluating the performance of proposed feature selection algorithms. To do this, the thorax surface region was divided into nine smaller segments and predefined tumors motion was predicted by ANFIS using external motion data of given markers at each small segment, separately. Our comparative results showed that all feature selection algorithms can reasonably select specific external markers from those segments where the root mean square error of the ANFIS model is minimum. Moreover, the performance accuracy of proposed feature selection algorithms was compared, separately. For this, each tumor motion was predicted using motion data of those external markers selected by each feature selection algorithm. Duncan statistical test, followed by F-test, on final results reflected that all proposed feature selection algorithms have the same performance accuracy for lung tumors. But for liver tumors, a correlation-based feature selection algorithm, in combination with a genetic search algorithm, proved to yield best performance accuracy for selecting optimum markers.

Borderline Personality Features in Students: the Predicting Role of Schema, Emotion Regulation, Dissociative Experience and Suicidal Ideation.

PubMed

Sajadi, Seyede Fateme; Arshadi, Nasrin; Zargar, Yadolla; Mehrabizade Honarmand, Mahnaz; Hajjari, Zahra

2015-06-01

Numerous studies have demonstrated that early maladaptive schemas, emotional dysregulation are supposed to be the defining core of borderline personality disorder. Many studies have also found a strong association between the diagnosis of borderline personality and the occurrence of suicide ideation and dissociative symptoms. The present study was designed to investigate the relationship between borderline personality features and schema, emotion regulation, dissociative experiences and suicidal ideation among high school students in Shiraz City, Iran. In this descriptive correlational study, 300 students (150 boys and 150 girls) were selected from the high schools in Shiraz, Iran, using the multi-stage random sampling. Data were collected using some instruments including borderline personality feature scale for children, young schema questionnaire-short form, difficulties in emotion-regulation scale (DERS), dissociative experience scale and beck suicide ideation scale. Data were analyzed using the Pearson correlation coefficient and multivariate regression analysis. The results showed a significant positive correlation between schema, emotion regulation, dissociative experiences and suicide ideation with borderline personality features. Moreover, the results of multivariate regression analysis suggested that among the studied variables, schema was the most effective predicting variable of borderline features (P < 0.001). The findings of this study are in accordance with findings from previous studies, and generally show a meaningful association between schema, emotion regulation, dissociative experiences, and suicide ideation with borderline personality features.

An Ant Colony Optimization Based Feature Selection for Web Page Classification

PubMed Central

2014-01-01

The increased popularity of the web has caused the inclusion of huge amount of information to the web, and as a result of this explosive information growth, automated web page classification systems are needed to improve search engines' performance. Web pages have a large number of features such as HTML/XML tags, URLs, hyperlinks, and text contents that should be considered during an automated classification process. The aim of this study is to reduce the number of features to be used to improve runtime and accuracy of the classification of web pages. In this study, we used an ant colony optimization (ACO) algorithm to select the best features, and then we applied the well-known C4.5, naive Bayes, and k nearest neighbor classifiers to assign class labels to web pages. We used the WebKB and Conference datasets in our experiments, and we showed that using the ACO for feature selection improves both accuracy and runtime performance of classification. We also showed that the proposed ACO based algorithm can select better features with respect to the well-known information gain and chi square feature selection methods. PMID:25136678

Hypergraph Based Feature Selection Technique for Medical Diagnosis.

PubMed

Somu, Nivethitha; Raman, M R Gauthama; Kirthivasan, Kannan; Sriram, V S Shankar

2016-11-01

The impact of internet and information systems across various domains have resulted in substantial generation of multidimensional datasets. The use of data mining and knowledge discovery techniques to extract the original information contained in the multidimensional datasets play a significant role in the exploitation of complete benefit provided by them. The presence of large number of features in the high dimensional datasets incurs high computational cost in terms of computing power and time. Hence, feature selection technique has been commonly used to build robust machine learning models to select a subset of relevant features which projects the maximal information content of the original dataset. In this paper, a novel Rough Set based K - Helly feature selection technique (RSKHT) which hybridize Rough Set Theory (RST) and K - Helly property of hypergraph representation had been designed to identify the optimal feature subset or reduct for medical diagnostic applications. Experiments carried out using the medical datasets from the UCI repository proves the dominance of the RSKHT over other feature selection techniques with respect to the reduct size, classification accuracy and time complexity. The performance of the RSKHT had been validated using WEKA tool, which shows that RSKHT had been computationally attractive and flexible over massive datasets.

Predicting infant cortical surface development using a 4D varifold-based learning framework and local topography-based shape morphing.

PubMed

Rekik, Islem; Li, Gang; Lin, Weili; Shen, Dinggang

2016-02-01

Longitudinal neuroimaging analysis methods have remarkably advanced our understanding of early postnatal brain development. However, learning predictive models to trace forth the evolution trajectories of both normal and abnormal cortical shapes remains broadly absent. To fill this critical gap, we pioneered the first prediction model for longitudinal developing cortical surfaces in infants using a spatiotemporal current-based learning framework solely from the baseline cortical surface. In this paper, we detail this prediction model and even further improve its performance by introducing two key variants. First, we use the varifold metric to overcome the limitations of the current metric for surface registration that was used in our preliminary study. We also extend the conventional varifold-based surface registration model for pairwise registration to a spatiotemporal surface regression model. Second, we propose a morphing process of the baseline surface using its topographic attributes such as normal direction and principal curvature sign. Specifically, our method learns from longitudinal data both the geometric (vertices positions) and dynamic (temporal evolution trajectories) features of the infant cortical surface, comprising a training stage and a prediction stage. In the training stage, we use the proposed varifold-based shape regression model to estimate geodesic cortical shape evolution trajectories for each training subject. We then build an empirical mean spatiotemporal surface atlas. In the prediction stage, given an infant, we select the best learnt features from training subjects to simultaneously predict the cortical surface shapes at all later timepoints, based on similarity metrics between this baseline surface and the learnt baseline population average surface atlas. We used a leave-one-out cross validation method to predict the inner cortical surface shape at 3, 6, 9 and 12 months of age from the baseline cortical surface shape at birth. Our method attained a higher prediction accuracy and better captured the spatiotemporal dynamic change of the highly folded cortical surface than the previous proposed prediction method. Copyright © 2015 Elsevier B.V. All rights reserved.

[Combining speech sample and feature bilateral selection algorithm for classification of Parkinson's disease].

PubMed

Zhang, Xiaoheng; Wang, Lirui; Cao, Yao; Wang, Pin; Zhang, Cheng; Yang, Liuyang; Li, Yongming; Zhang, Yanling; Cheng, Oumei

2018-02-01

Diagnosis of Parkinson's disease (PD) based on speech data has been proved to be an effective way in recent years. However, current researches just care about the feature extraction and classifier design, and do not consider the instance selection. Former research by authors showed that the instance selection can lead to improvement on classification accuracy. However, no attention is paid on the relationship between speech sample and feature until now. Therefore, a new diagnosis algorithm of PD is proposed in this paper by simultaneously selecting speech sample and feature based on relevant feature weighting algorithm and multiple kernel method, so as to find their synergy effects, thereby improving classification accuracy. Experimental results showed that this proposed algorithm obtained apparent improvement on classification accuracy. It can obtain mean classification accuracy of 82.5%, which was 30.5% higher than the relevant algorithm. Besides, the proposed algorithm detected the synergy effects of speech sample and feature, which is valuable for speech marker extraction.

Mapping Regional Impervious Surface Distribution from Night Time Light: The Variability across Global Cities

NASA Astrophysics Data System (ADS)

Lin, M.; Yang, Z.; Park, H.; Qian, S.; Chen, J.; Fan, P.

2017-12-01

Impervious surface area (ISA) has become an important indicator for studying urban environments, but mapping ISA at the regional or global scale is still challenging due to the complexity of impervious surface features. The Defense Meteorological Satellite Program's Operational Linescan System (DMSP-OLS) nighttime light data is (NTL) and Resolution Imaging Spectroradiometer (MODIS) are the major remote sensing data source for regional ISA mapping. A single regression relationship between fractional ISA and NTL or various index derived based on NTL and MODIS vegetation index (NDVI) data was established in many previous studies for regional ISA mapping. However, due to the varying geographical, climatic, and socio-economic characteristics of different cities, the same regression relationship may vary significantly across different cities in the same region in terms of both fitting performance (i.e. R2) and the rate of change (Slope). In this study, we examined the regression relationship between fractional ISA and Vegetation Adjusted Nighttime light Urban Index (VANUI) for 120 randomly selected cities around the world with a multilevel regression model. We found that indeed there is substantial variability of both the R2 (0.68±0.29) and slopes (0.64±0.40) among individual regressions, which suggests that multilevel/hierarchical models are needed for accuracy improvement of future regional ISA mapping .Further analysis also let us find the this substantial variability are affected by climate conditions, socio-economic status, and urban spatial structures. However, all these effects are nonlinear rather than linear, thus could not modeled explicitly in multilevel linear regression models.

Selection of nest-site habitat by interior least terns in relation to sandbar construction

USGS Publications Warehouse

Sherfy, M.H.; Stucker, J.H.; Buhl, D.A.

2012-01-01

Federally endangered interior least terns (Sternula antillarum) nest on bare or sparsely vegetated sandbars on midcontinent river systems. Loss of nesting habitat has been implicated as a cause of population declines, and managing these habitats is a major initiative in population recovery. One such initiative involves construction of mid-channel sandbars on the Missouri River, where natural sandbar habitat has declined in quantity and quality since the late 1990s. We evaluated nest-site habitat selection by least terns on constructed and natural sandbars by comparing vegetation, substrate, and debris variables at nest sites (na =a 798) and random points (na =a 1,113) in bare or sparsely vegetated habitats. Our logistic regression models revealed that a broader suite of habitat features was important in nest-site selection on constructed than on natural sandbars. Odds ratios for habitat variables indicated that avoidance of habitat features was the dominant nest-site selection process on both sandbar types, with nesting terns being attracted to nest-site habitat features (gravel and debris) and avoiding vegetation only on constructed sandbars, and avoiding silt and leaf litter on both sandbar types. Despite the seemingly uniform nature of these habitats, our results suggest that a complex suite of habitat features influences nest-site choice by least terns. However, nest-site selection in this social, colonially nesting species may be influenced by other factors, including spatial arrangement of bare sand habitat, proximity to other least terns, and prior habitat occupancy by piping plovers (Charadrius melodus). We found that nest-site selection was sensitive to subtle variation in habitat features, suggesting that rigor in maintaining habitat condition will be necessary in managing sandbars for the benefit of least terns. Further, management strategies that reduce habitat features that are avoided by least terns may be the most beneficial to nesting least terns. ?? 2011 The Wildlife Society.

Selection of nest-site habitat by interior least terns in relation to sandbar construction

USGS Publications Warehouse

Sherfy, Mark H.; Stucker, Jennifer H.; Buhl, Deborah A.

2012-01-01

Federally endangered interior least terns (Sternula antillarum) nest on bare or sparsely vegetated sandbars on midcontinent river systems. Loss of nesting habitat has been implicated as a cause of population declines, and managing these habitats is a major initiative in population recovery. One such initiative involves construction of mid-channel sandbars on the Missouri River, where natural sandbar habitat has declined in quantity and quality since the late 1990s. We evaluated nest-site habitat selection by least terns on constructed and natural sandbars by comparing vegetation, substrate, and debris variables at nest sites (n = 798) and random points (n = 1,113) in bare or sparsely vegetated habitats. Our logistic regression models revealed that a broader suite of habitat features was important in nest-site selection on constructed than on natural sandbars. Odds ratios for habitat variables indicated that avoidance of habitat features was the dominant nest-site selection process on both sandbar types, with nesting terns being attracted to nest-site habitat features (gravel and debris) and avoiding vegetation only on constructed sandbars, and avoiding silt and leaf litter on both sandbar types. Despite the seemingly uniform nature of these habitats, our results suggest that a complex suite of habitat features influences nest-site choice by least terns. However, nest-site selection in this social, colonially nesting species may be influenced by other factors, including spatial arrangement of bare sand habitat, proximity to other least terns, and prior habitat occupancy by piping plovers (Charadrius melodus). We found that nest-site selection was sensitive to subtle variation in habitat features, suggesting that rigor in maintaining habitat condition will be necessary in managing sandbars for the benefit of least terns. Further, management strategies that reduce habitat features that are avoided by least terns may be the most beneficial to nesting least terns.

Image search engine with selective filtering and feature-element-based classification

NASA Astrophysics Data System (ADS)

Li, Qing; Zhang, Yujin; Dai, Shengyang

2001-12-01

With the growth of Internet and storage capability in recent years, image has become a widespread information format in World Wide Web. However, it has become increasingly harder to search for images of interest, and effective image search engine for the WWW needs to be developed. We propose in this paper a selective filtering process and a novel approach for image classification based on feature element in the image search engine we developed for the WWW. First a selective filtering process is embedded in a general web crawler to filter out the meaningless images with GIF format. Two parameters that can be obtained easily are used in the filtering process. Our classification approach first extract feature elements from images instead of feature vectors. Compared with feature vectors, feature elements can better capture visual meanings of the image according to subjective perception of human beings. Different from traditional image classification method, our classification approach based on feature element doesn't calculate the distance between two vectors in the feature space, while trying to find associations between feature element and class attribute of the image. Experiments are presented to show the efficiency of the proposed approach.

Regression estimators for generic health-related quality of life and quality-adjusted life years.

PubMed

Basu, Anirban; Manca, Andrea

2012-01-01

To develop regression models for outcomes with truncated supports, such as health-related quality of life (HRQoL) data, and account for features typical of such data such as a skewed distribution, spikes at 1 or 0, and heteroskedasticity. Regression estimators based on features of the Beta distribution. First, both a single equation and a 2-part model are presented, along with estimation algorithms based on maximum-likelihood, quasi-likelihood, and Bayesian Markov-chain Monte Carlo methods. A novel Bayesian quasi-likelihood estimator is proposed. Second, a simulation exercise is presented to assess the performance of the proposed estimators against ordinary least squares (OLS) regression for a variety of HRQoL distributions that are encountered in practice. Finally, the performance of the proposed estimators is assessed by using them to quantify the treatment effect on QALYs in the EVALUATE hysterectomy trial. Overall model fit is studied using several goodness-of-fit tests such as Pearson's correlation test, link and reset tests, and a modified Hosmer-Lemeshow test. The simulation results indicate that the proposed methods are more robust in estimating covariate effects than OLS, especially when the effects are large or the HRQoL distribution has a large spike at 1. Quasi-likelihood techniques are more robust than maximum likelihood estimators. When applied to the EVALUATE trial, all but the maximum likelihood estimators produce unbiased estimates of the treatment effect. One and 2-part Beta regression models provide flexible approaches to regress the outcomes with truncated supports, such as HRQoL, on covariates, after accounting for many idiosyncratic features of the outcomes distribution. This work will provide applied researchers with a practical set of tools to model outcomes in cost-effectiveness analysis.

Influence of time and length size feature selections for human activity sequences recognition.

PubMed

Fang, Hongqing; Chen, Long; Srinivasan, Raghavendiran

2014-01-01

In this paper, Viterbi algorithm based on a hidden Markov model is applied to recognize activity sequences from observed sensors events. Alternative features selections of time feature values of sensors events and activity length size feature values are tested, respectively, and then the results of activity sequences recognition performances of Viterbi algorithm are evaluated. The results show that the selection of larger time feature values of sensor events and/or smaller activity length size feature values will generate relatively better results on the activity sequences recognition performances. © 2013 ISA Published by ISA All rights reserved.

Using GA-Ridge regression to select hydro-geological parameters influencing groundwater pollution vulnerability.

PubMed

Ahn, Jae Joon; Kim, Young Min; Yoo, Keunje; Park, Joonhong; Oh, Kyong Joo

2012-11-01

For groundwater conservation and management, it is important to accurately assess groundwater pollution vulnerability. This study proposed an integrated model using ridge regression and a genetic algorithm (GA) to effectively select the major hydro-geological parameters influencing groundwater pollution vulnerability in an aquifer. The GA-Ridge regression method determined that depth to water, net recharge, topography, and the impact of vadose zone media were the hydro-geological parameters that influenced trichloroethene pollution vulnerability in a Korean aquifer. When using these selected hydro-geological parameters, the accuracy was improved for various statistical nonlinear and artificial intelligence (AI) techniques, such as multinomial logistic regression, decision trees, artificial neural networks, and case-based reasoning. These results provide a proof of concept that the GA-Ridge regression is effective at determining influential hydro-geological parameters for the pollution vulnerability of an aquifer, and in turn, improves the AI performance in assessing groundwater pollution vulnerability.

Integrated Low-Rank-Based Discriminative Feature Learning for Recognition.

PubMed

Zhou, Pan; Lin, Zhouchen; Zhang, Chao

2016-05-01

Feature learning plays a central role in pattern recognition. In recent years, many representation-based feature learning methods have been proposed and have achieved great success in many applications. However, these methods perform feature learning and subsequent classification in two separate steps, which may not be optimal for recognition tasks. In this paper, we present a supervised low-rank-based approach for learning discriminative features. By integrating latent low-rank representation (LatLRR) with a ridge regression-based classifier, our approach combines feature learning with classification, so that the regulated classification error is minimized. In this way, the extracted features are more discriminative for the recognition tasks. Our approach benefits from a recent discovery on the closed-form solutions to noiseless LatLRR. When there is noise, a robust Principal Component Analysis (PCA)-based denoising step can be added as preprocessing. When the scale of a problem is large, we utilize a fast randomized algorithm to speed up the computation of robust PCA. Extensive experimental results demonstrate the effectiveness and robustness of our method.

Two-Year versus One-Year Head Start Program Impact: Addressing Selection Bias by Comparing Regression Modeling with Propensity Score Analysis

ERIC Educational Resources Information Center

Leow, Christine; Wen, Xiaoli; Korfmacher, Jon

2015-01-01

This article compares regression modeling and propensity score analysis as different types of statistical techniques used in addressing selection bias when estimating the impact of two-year versus one-year Head Start on children's school readiness. The analyses were based on the national Head Start secondary dataset. After controlling for…

The Speed of Feature-Based Attention: Attentional Advantage Is Slow, but Selection Is Fast

ERIC Educational Resources Information Center

Huang, Liqiang

2010-01-01

When paying attention to a feature (e.g., red), no attentional advantage is gained in perceiving items with this feature in very brief displays. Therefore, feature-based attention seems to be slow. In previous feature-based attention studies, attention has often been measured as the difference in performance in a secondary task. In our recent work…

Application-Dedicated Selection of Filters (ADSF) using covariance maximization and orthogonal projection.

PubMed

Hadoux, Xavier; Kumar, Dinesh Kant; Sarossy, Marc G; Roger, Jean-Michel; Gorretta, Nathalie

2016-05-19

Visible and near-infrared (Vis-NIR) spectra are generated by the combination of numerous low resolution features. Spectral variables are thus highly correlated, which can cause problems for selecting the most appropriate ones for a given application. Some decomposition bases such as Fourier or wavelet generally help highlighting spectral features that are important, but are by nature constraint to have both positive and negative components. Thus, in addition to complicating the selected features interpretability, it impedes their use for application-dedicated sensors. In this paper we have proposed a new method for feature selection: Application-Dedicated Selection of Filters (ADSF). This method relaxes the shape constraint by enabling the selection of any type of user defined custom features. By considering only relevant features, based on the underlying nature of the data, high regularization of the final model can be obtained, even in the small sample size context often encountered in spectroscopic applications. For larger scale deployment of application-dedicated sensors, these predefined feature constraints can lead to application specific optical filters, e.g., lowpass, highpass, bandpass or bandstop filters with positive only coefficients. In a similar fashion to Partial Least Squares, ADSF successively selects features using covariance maximization and deflates their influences using orthogonal projection in order to optimally tune the selection to the data with limited redundancy. ADSF is well suited for spectroscopic data as it can deal with large numbers of highly correlated variables in supervised learning, even with many correlated responses. Copyright © 2016 Elsevier B.V. All rights reserved.

Predicting individualized clinical measures by a generalized prediction framework and multimodal fusion of MRI data

PubMed Central

Meng, Xing; Jiang, Rongtao; Lin, Dongdong; Bustillo, Juan; Jones, Thomas; Chen, Jiayu; Yu, Qingbao; Du, Yuhui; Zhang, Yu; Jiang, Tianzi; Sui, Jing; Calhoun, Vince D.

2016-01-01

Neuroimaging techniques have greatly enhanced the understanding of neurodiversity (human brain variation across individuals) in both health and disease. The ultimate goal of using brain imaging biomarkers is to perform individualized predictions. Here we proposed a generalized framework that can predict explicit values of the targeted measures by taking advantage of joint information from multiple modalities. This framework also enables whole brain voxel-wise searching by combining multivariate techniques such as ReliefF, clustering, correlation-based feature selection and multiple regression models, which is more flexible and can achieve better prediction performance than alternative atlas-based methods. For 50 healthy controls and 47 schizophrenia patients, three kinds of features derived from resting-state fMRI (fALFF), sMRI (gray matter) and DTI (fractional anisotropy) were extracted and fed into a regression model, achieving high prediction for both cognitive scores (MCCB composite r = 0.7033, MCCB social cognition r = 0.7084) and symptomatic scores (positive and negative syndrome scale [PANSS] positive r = 0.7785, PANSS negative r = 0.7804). Moreover, the brain areas likely responsible for cognitive deficits of schizophrenia, including middle temporal gyrus, dorsolateral prefrontal cortex, striatum, cuneus and cerebellum, were located with different weights, as well as regions predicting PANSS symptoms, including thalamus, striatum and inferior parietal lobule, pinpointing the potential neuromarkers. Finally, compared to a single modality, multimodal combination achieves higher prediction accuracy and enables individualized prediction on multiple clinical measures. There is more work to be done, but the current results highlight the potential utility of multimodal brain imaging biomarkers to eventually inform clinical decision-making. PMID:27177764

A multi-feature integration method for fatigue crack detection and crack length estimation in riveted lap joints using Lamb waves

NASA Astrophysics Data System (ADS)

He, Jingjing; Guan, Xuefei; Peng, Tishun; Liu, Yongming; Saxena, Abhinav; Celaya, Jose; Goebel, Kai

2013-10-01

This paper presents an experimental study of damage detection and quantification in riveted lap joints. Embedded lead zirconate titanate piezoelectric (PZT) ceramic wafer-type sensors are employed to perform in situ non-destructive evaluation (NDE) during fatigue cyclical loading. PZT wafers are used to monitor the wave reflection from the boundaries of the fatigue crack at the edge of bolt joints. The group velocity of the guided wave is calculated to select a proper time window in which the received signal contains the damage information. It is found that the fatigue crack lengths are correlated with three main features of the signal, i.e., correlation coefficient, amplitude change, and phase change. It was also observed that a single feature cannot be used to quantify the damage among different specimens since a considerable variability was observed in the response from different specimens. A multi-feature integration method based on a second-order multivariate regression analysis is proposed for the prediction of fatigue crack lengths using sensor measurements. The model parameters are obtained using training datasets from five specimens. The effectiveness of the proposed methodology is demonstrated using several lap joint specimens from different manufactures and under different loading conditions.

Quantitative Analysis of the Cervical Texture by Ultrasound and Correlation with Gestational Age.

PubMed

Baños, Núria; Perez-Moreno, Alvaro; Migliorelli, Federico; Triginer, Laura; Cobo, Teresa; Bonet-Carne, Elisenda; Gratacos, Eduard; Palacio, Montse

2017-01-01

Quantitative texture analysis has been proposed to extract robust features from the ultrasound image to detect subtle changes in the textures of the images. The aim of this study was to evaluate the feasibility of quantitative cervical texture analysis to assess cervical tissue changes throughout pregnancy. This was a cross-sectional study including singleton pregnancies between 20.0 and 41.6 weeks of gestation from women who delivered at term. Cervical length was measured, and a selected region of interest in the cervix was delineated. A model to predict gestational age based on features extracted from cervical images was developed following three steps: data splitting, feature transformation, and regression model computation. Seven hundred images, 30 per gestational week, were included for analysis. There was a strong correlation between the gestational age at which the images were obtained and the estimated gestational age by quantitative analysis of the cervical texture (R = 0.88). This study provides evidence that quantitative analysis of cervical texture can extract features from cervical ultrasound images which correlate with gestational age. Further research is needed to evaluate its applicability as a biomarker of the risk of spontaneous preterm birth, as well as its role in cervical assessment in other clinical situations in which cervical evaluation might be relevant. © 2016 S. Karger AG, Basel.

Efficient Iris Recognition Based on Optimal Subfeature Selection and Weighted Subregion Fusion

PubMed Central

Deng, Ning

2014-01-01

In this paper, we propose three discriminative feature selection strategies and weighted subregion matching method to improve the performance of iris recognition system. Firstly, we introduce the process of feature extraction and representation based on scale invariant feature transformation (SIFT) in detail. Secondly, three strategies are described, which are orientation probability distribution function (OPDF) based strategy to delete some redundant feature keypoints, magnitude probability distribution function (MPDF) based strategy to reduce dimensionality of feature element, and compounded strategy combined OPDF and MPDF to further select optimal subfeature. Thirdly, to make matching more effective, this paper proposes a novel matching method based on weighted sub-region matching fusion. Particle swarm optimization is utilized to accelerate achieve different sub-region's weights and then weighted different subregions' matching scores to generate the final decision. The experimental results, on three public and renowned iris databases (CASIA-V3 Interval, Lamp, andMMU-V1), demonstrate that our proposed methods outperform some of the existing methods in terms of correct recognition rate, equal error rate, and computation complexity. PMID:24683317

Efficient iris recognition based on optimal subfeature selection and weighted subregion fusion.

PubMed

Chen, Ying; Liu, Yuanning; Zhu, Xiaodong; He, Fei; Wang, Hongye; Deng, Ning

2014-01-01

In this paper, we propose three discriminative feature selection strategies and weighted subregion matching method to improve the performance of iris recognition system. Firstly, we introduce the process of feature extraction and representation based on scale invariant feature transformation (SIFT) in detail. Secondly, three strategies are described, which are orientation probability distribution function (OPDF) based strategy to delete some redundant feature keypoints, magnitude probability distribution function (MPDF) based strategy to reduce dimensionality of feature element, and compounded strategy combined OPDF and MPDF to further select optimal subfeature. Thirdly, to make matching more effective, this paper proposes a novel matching method based on weighted sub-region matching fusion. Particle swarm optimization is utilized to accelerate achieve different sub-region's weights and then weighted different subregions' matching scores to generate the final decision. The experimental results, on three public and renowned iris databases (CASIA-V3 Interval, Lamp, and MMU-V1), demonstrate that our proposed methods outperform some of the existing methods in terms of correct recognition rate, equal error rate, and computation complexity.

TANDEM: a two-stage approach to maximize interpretability of drug response models based on multiple molecular data types.

PubMed

Aben, Nanne; Vis, Daniel J; Michaut, Magali; Wessels, Lodewyk F A

2016-09-01

Clinical response to anti-cancer drugs varies between patients. A large portion of this variation can be explained by differences in molecular features, such as mutation status, copy number alterations, methylation and gene expression profiles. We show that the classic approach for combining these molecular features (Elastic Net regression on all molecular features simultaneously) results in models that are almost exclusively based on gene expression. The gene expression features selected by the classic approach are difficult to interpret as they often represent poorly studied combinations of genes, activated by aberrations in upstream signaling pathways. To utilize all data types in a more balanced way, we developed TANDEM, a two-stage approach in which the first stage explains response using upstream features (mutations, copy number, methylation and cancer type) and the second stage explains the remainder using downstream features (gene expression). Applying TANDEM to 934 cell lines profiled across 265 drugs (GDSC1000), we show that the resulting models are more interpretable, while retaining the same predictive performance as the classic approach. Using the more balanced contributions per data type as determined with TANDEM, we find that response to MAPK pathway inhibitors is largely predicted by mutation data, while predicting response to DNA damaging agents requires gene expression data, in particular SLFN11 expression. TANDEM is available as an R package on CRAN (for more information, see http://ccb.nki.nl/software/tandem). m.michaut@nki.nl or l.wessels@nki.nl Supplementary data are available at Bioinformatics online. © The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.

Drug-target interaction prediction using ensemble learning and dimensionality reduction.

PubMed

Ezzat, Ali; Wu, Min; Li, Xiao-Li; Kwoh, Chee-Keong

2017-10-01

Experimental prediction of drug-target interactions is expensive, time-consuming and tedious. Fortunately, computational methods help narrow down the search space for interaction candidates to be further examined via wet-lab techniques. Nowadays, the number of attributes/features for drugs and targets, as well as the amount of their interactions, are increasing, making these computational methods inefficient or occasionally prohibitive. This motivates us to derive a reduced feature set for prediction. In addition, since ensemble learning techniques are widely used to improve the classification performance, it is also worthwhile to design an ensemble learning framework to enhance the performance for drug-target interaction prediction. In this paper, we propose a framework for drug-target interaction prediction leveraging both feature dimensionality reduction and ensemble learning. First, we conducted feature subspacing to inject diversity into the classifier ensemble. Second, we applied three different dimensionality reduction methods to the subspaced features. Third, we trained homogeneous base learners with the reduced features and then aggregated their scores to derive the final predictions. For base learners, we selected two classifiers, namely Decision Tree and Kernel Ridge Regression, resulting in two variants of ensemble models, EnsemDT and EnsemKRR, respectively. In our experiments, we utilized AUC (Area under ROC Curve) as an evaluation metric. We compared our proposed methods with various state-of-the-art methods under 5-fold cross validation. Experimental results showed EnsemKRR achieving the highest AUC (94.3%) for predicting drug-target interactions. In addition, dimensionality reduction helped improve the performance of EnsemDT. In conclusion, our proposed methods produced significant improvements for drug-target interaction prediction. Copyright © 2017 Elsevier Inc. All rights reserved.

A Practical Guide to Regression Discontinuity

ERIC Educational Resources Information Center

Jacob, Robin; Zhu, Pei; Somers, Marie-Andrée; Bloom, Howard

2012-01-01

Regression discontinuity (RD) analysis is a rigorous nonexperimental approach that can be used to estimate program impacts in situations in which candidates are selected for treatment based on whether their value for a numeric rating exceeds a designated threshold or cut-point. Over the last two decades, the regression discontinuity approach has…

Unbiased feature selection in learning random forests for high-dimensional data.

PubMed

Nguyen, Thanh-Tung; Huang, Joshua Zhexue; Nguyen, Thuy Thi

2015-01-01

Random forests (RFs) have been widely used as a powerful classification method. However, with the randomization in both bagging samples and feature selection, the trees in the forest tend to select uninformative features for node splitting. This makes RFs have poor accuracy when working with high-dimensional data. Besides that, RFs have bias in the feature selection process where multivalued features are favored. Aiming at debiasing feature selection in RFs, we propose a new RF algorithm, called xRF, to select good features in learning RFs for high-dimensional data. We first remove the uninformative features using p-value assessment, and the subset of unbiased features is then selected based on some statistical measures. This feature subset is then partitioned into two subsets. A feature weighting sampling technique is used to sample features from these two subsets for building trees. This approach enables one to generate more accurate trees, while allowing one to reduce dimensionality and the amount of data needed for learning RFs. An extensive set of experiments has been conducted on 47 high-dimensional real-world datasets including image datasets. The experimental results have shown that RFs with the proposed approach outperformed the existing random forests in increasing the accuracy and the AUC measures.

Using multiple classifiers for predicting the risk of endovascular aortic aneurysm repair re-intervention through hybrid feature selection.

PubMed

Attallah, Omneya; Karthikesalingam, Alan; Holt, Peter Je; Thompson, Matthew M; Sayers, Rob; Bown, Matthew J; Choke, Eddie C; Ma, Xianghong

2017-11-01

Feature selection is essential in medical area; however, its process becomes complicated with the presence of censoring which is the unique character of survival analysis. Most survival feature selection methods are based on Cox's proportional hazard model, though machine learning classifiers are preferred. They are less employed in survival analysis due to censoring which prevents them from directly being used to survival data. Among the few work that employed machine learning classifiers, partial logistic artificial neural network with auto-relevance determination is a well-known method that deals with censoring and perform feature selection for survival data. However, it depends on data replication to handle censoring which leads to unbalanced and biased prediction results especially in highly censored data. Other methods cannot deal with high censoring. Therefore, in this article, a new hybrid feature selection method is proposed which presents a solution to high level censoring. It combines support vector machine, neural network, and K-nearest neighbor classifiers using simple majority voting and a new weighted majority voting method based on survival metric to construct a multiple classifier system. The new hybrid feature selection process uses multiple classifier system as a wrapper method and merges it with iterated feature ranking filter method to further reduce features. Two endovascular aortic repair datasets containing 91% censored patients collected from two centers were used to construct a multicenter study to evaluate the performance of the proposed approach. The results showed the proposed technique outperformed individual classifiers and variable selection methods based on Cox's model such as Akaike and Bayesian information criterions and least absolute shrinkage and selector operator in p values of the log-rank test, sensitivity, and concordance index. This indicates that the proposed classifier is more powerful in correctly predicting the risk of re-intervention enabling doctor in selecting patients' future follow-up plan.

Adaptive feature selection using v-shaped binary particle swarm optimization.

PubMed

Teng, Xuyang; Dong, Hongbin; Zhou, Xiurong

2017-01-01

Feature selection is an important preprocessing method in machine learning and data mining. This process can be used not only to reduce the amount of data to be analyzed but also to build models with stronger interpretability based on fewer features. Traditional feature selection methods evaluate the dependency and redundancy of features separately, which leads to a lack of measurement of their combined effect. Moreover, a greedy search considers only the optimization of the current round and thus cannot be a global search. To evaluate the combined effect of different subsets in the entire feature space, an adaptive feature selection method based on V-shaped binary particle swarm optimization is proposed. In this method, the fitness function is constructed using the correlation information entropy. Feature subsets are regarded as individuals in a population, and the feature space is searched using V-shaped binary particle swarm optimization. The above procedure overcomes the hard constraint on the number of features, enables the combined evaluation of each subset as a whole, and improves the search ability of conventional binary particle swarm optimization. The proposed algorithm is an adaptive method with respect to the number of feature subsets. The experimental results show the advantages of optimizing the feature subsets using the V-shaped transfer function and confirm the effectiveness and efficiency of the feature subsets obtained under different classifiers.

Adaptive feature selection using v-shaped binary particle swarm optimization

PubMed Central

Dong, Hongbin; Zhou, Xiurong

2017-01-01

Feature selection is an important preprocessing method in machine learning and data mining. This process can be used not only to reduce the amount of data to be analyzed but also to build models with stronger interpretability based on fewer features. Traditional feature selection methods evaluate the dependency and redundancy of features separately, which leads to a lack of measurement of their combined effect. Moreover, a greedy search considers only the optimization of the current round and thus cannot be a global search. To evaluate the combined effect of different subsets in the entire feature space, an adaptive feature selection method based on V-shaped binary particle swarm optimization is proposed. In this method, the fitness function is constructed using the correlation information entropy. Feature subsets are regarded as individuals in a population, and the feature space is searched using V-shaped binary particle swarm optimization. The above procedure overcomes the hard constraint on the number of features, enables the combined evaluation of each subset as a whole, and improves the search ability of conventional binary particle swarm optimization. The proposed algorithm is an adaptive method with respect to the number of feature subsets. The experimental results show the advantages of optimizing the feature subsets using the V-shaped transfer function and confirm the effectiveness and efficiency of the feature subsets obtained under different classifiers. PMID:28358850

Investigating a memory-based account of negative priming: support for selection-feature mismatch.

PubMed

MacDonald, P A; Joordens, S

2000-08-01

Using typical and modified negative priming tasks, the selection-feature mismatch account of negative priming was tested. In the modified task, participants performed selections on the basis of a semantic feature (e.g., referent size). This procedure has been shown to enhance negative priming (P. A. MacDonald, S. Joordens, & K. N. Seergobin, 1999). Across 3 experiments, negative priming occurred only when the repeated item mismatched in terms of the feature used as the basis for selections. When the repeated item was congruent on the selection feature across the prime and probe displays, positive priming arose. This pattern of results appeared in both the ignored- and the attended-repetition conditions. Negative priming does not result from previously ignoring an item. These findings strongly support the selection-feature mismatch account of negative priming and refute both the distractor inhibition and the episodic-retrieval explanations.

Generic Feature Selection with Short Fat Data

PubMed Central

Clarke, B.; Chu, J.-H.

2014-01-01

SUMMARY Consider a regression problem in which there are many more explanatory variables than data points, i.e., p ≫ n. Essentially, without reducing the number of variables inference is impossible. So, we group the p explanatory variables into blocks by clustering, evaluate statistics on the blocks and then regress the response on these statistics under a penalized error criterion to obtain estimates of the regression coefficients. We examine the performance of this approach for a variety of choices of n, p, classes of statistics, clustering algorithms, penalty terms, and data types. When n is not large, the discrimination over number of statistics is weak, but computations suggest regressing on approximately [n/K] statistics where K is the number of blocks formed by a clustering algorithm. Small deviations from this are observed when the blocks of variables are of very different sizes. Larger deviations are observed when the penalty term is an Lq norm with high enough q. PMID:25346546

Hyperspectral Features of Oil-Polluted Sea Ice and the Response to the Contamination Area Fraction

PubMed Central

Li, Ying; Liu, Chengyu; Xie, Feng

2018-01-01

Researchers have studied oil spills in open waters using remote sensors, but few have focused on extracting reflectance features of oil pollution on sea ice. An experiment was conducted on natural sea ice in Bohai Bay, China, to obtain the spectral reflectance of oil-contaminated sea ice. The spectral absorption index (SAI), spectral peak height (SPH), and wavelet detail coefficient (DWT d5) were calculated using stepwise multiple linear regression. The reflectances of some false targets were measured and analysed. The simulated false targets were sediment, iron ore fines, coal dust, and the melt pool. The measured reflectances were resampled using five common sensors (GF-2, Landsat8-OLI, Sentinel3-OLCI, MODIS, and AVIRIS). Some significant spectral features could discriminate between oil-polluted and clean sea ice. The indices correlated well with the oil area fractions. All of the adjusted R2 values exceeded 0.9. The SPH model1, based on spectral features at 507–670 and 1627–1746 nm, displayed the best fitting. The resampled data indicated that these multi-spectral and hyper-spectral sensors could be used to detect crude oil on the sea ice if the effect of noise and spatial resolution are neglected. The spectral features and their identified changes may provide reference on sensor design and band selection. PMID:29342945

4D-LQTA-QSAR and docking study on potent Gram-negative specific LpxC inhibitors: a comparison to CoMFA modeling.

PubMed

Ghasemi, Jahan B; Safavi-Sohi, Reihaneh; Barbosa, Euzébio G

2012-02-01

A quasi 4D-QSAR has been carried out on a series of potent Gram-negative LpxC inhibitors. This approach makes use of the molecular dynamics (MD) trajectories and topology information retrieved from the GROMACS package. This new methodology is based on the generation of a conformational ensemble profile, CEP, for each compound instead of only one conformation, followed by the calculation intermolecular interaction energies at each grid point considering probes and all aligned conformations resulting from MD simulations. These interaction energies are independent variables employed in a QSAR analysis. The comparison of the proposed methodology to comparative molecular field analysis (CoMFA) formalism was performed. This methodology explores jointly the main features of CoMFA and 4D-QSAR models. Step-wise multiple linear regression was used for the selection of the most informative variables. After variable selection, multiple linear regression (MLR) and partial least squares (PLS) methods used for building the regression models. Leave-N-out cross-validation (LNO), and Y-randomization were performed in order to confirm the robustness of the model in addition to analysis of the independent test set. Best models provided the following statistics: [Formula in text] (PLS) and [Formula in text] (MLR). Docking study was applied to investigate the major interactions in protein-ligand complex with CDOCKER algorithm. Visualization of the descriptors of the best model helps us to interpret the model from the chemical point of view, supporting the applicability of this new approach in rational drug design.

Wrapper-based selection of genetic features in genome-wide association studies through fast matrix operations

PubMed Central

2012-01-01

Background Through the wealth of information contained within them, genome-wide association studies (GWAS) have the potential to provide researchers with a systematic means of associating genetic variants with a wide variety of disease phenotypes. Due to the limitations of approaches that have analyzed single variants one at a time, it has been proposed that the genetic basis of these disorders could be determined through detailed analysis of the genetic variants themselves and in conjunction with one another. The construction of models that account for these subsets of variants requires methodologies that generate predictions based on the total risk of a particular group of polymorphisms. However, due to the excessive number of variants, constructing these types of models has so far been computationally infeasible. Results We have implemented an algorithm, known as greedy RLS, that we use to perform the first known wrapper-based feature selection on the genome-wide level. The running time of greedy RLS grows linearly in the number of training examples, the number of features in the original data set, and the number of selected features. This speed is achieved through computational short-cuts based on matrix calculus. Since the memory consumption in present-day computers can form an even tighter bottleneck than running time, we also developed a space efficient variation of greedy RLS which trades running time for memory. These approaches are then compared to traditional wrapper-based feature selection implementations based on support vector machines (SVM) to reveal the relative speed-up and to assess the feasibility of the new algorithm. As a proof of concept, we apply greedy RLS to the Hypertension – UK National Blood Service WTCCC dataset and select the most predictive variants using 3-fold external cross-validation in less than 26 minutes on a high-end desktop. On this dataset, we also show that greedy RLS has a better classification performance on independent test data than a classifier trained using features selected by a statistical p-value-based filter, which is currently the most popular approach for constructing predictive models in GWAS. Conclusions Greedy RLS is the first known implementation of a machine learning based method with the capability to conduct a wrapper-based feature selection on an entire GWAS containing several thousand examples and over 400,000 variants. In our experiments, greedy RLS selected a highly predictive subset of genetic variants in a fraction of the time spent by wrapper-based selection methods used together with SVM classifiers. The proposed algorithms are freely available as part of the RLScore software library at http://users.utu.fi/aatapa/RLScore/. PMID:22551170

Forest Type and Above Ground Biomass Estimation Based on Sentinel-2A and WorldView-2 Data Evaluation of Predictor nd Data Suitability

NASA Astrophysics Data System (ADS)

Fritz, Andreas; Enßle, Fabian; Zhang, Xiaoli; Koch, Barbara

2016-08-01

The present study analyses the two earth observation sensors regarding their capability of modelling forest above ground biomass and forest density. Our research is carried out at two different demonstration sites. The first is located in south-western Germany (region Karlsruhe) and the second is located in southern China in Jiangle County (Province Fujian). A set of spectral and spatial predictors are computed from both, Sentinel-2A and WorldView-2 data. Window sizes in the range of 3*3 pixels to 21*21 pixels are computed in order to cover the full range of the canopy sizes of mature forest stands. Textural predictors of first and second order (grey-level-co-occurrence matrix) are calculated and are further used within a feature selection procedure. Additionally common spectral predictors from WorldView-2 and Sentinel-2A data such as all relevant spectral bands and NDVI are integrated in the analyses. To examine the most important predictors, a predictor selection algorithm is applied to the data, whereas the entire predictor set of more than 1000 predictors is used to find most important ones. Out of the original set only the most important predictors are then further analysed. Predictor selection is done with the Boruta package in R (Kursa and Rudnicki (2010)), whereas regression is computed with random forest. Prior the classification and regression a tuning of parameters is done by a repetitive model selection (100 runs), based on the .632 bootstrapping. Both are implemented in the caret R pack- age (Kuhn et al. (2016)). To account for the variability in the data set 100 independent runs are performed. Within each run 80 percent of the data is used for training and the 20 percent are used for an independent validation. With the subset of original predictors mapping of above ground biomass is performed.

Particle Swarm Optimization Based Feature Enhancement and Feature Selection for Improved Emotion Recognition in Speech and Glottal Signals

PubMed Central

Muthusamy, Hariharan; Polat, Kemal; Yaacob, Sazali

2015-01-01

In the recent years, many research works have been published using speech related features for speech emotion recognition, however, recent studies show that there is a strong correlation between emotional states and glottal features. In this work, Mel-frequency cepstralcoefficients (MFCCs), linear predictive cepstral coefficients (LPCCs), perceptual linear predictive (PLP) features, gammatone filter outputs, timbral texture features, stationary wavelet transform based timbral texture features and relative wavelet packet energy and entropy features were extracted from the emotional speech (ES) signals and its glottal waveforms(GW). Particle swarm optimization based clustering (PSOC) and wrapper based particle swarm optimization (WPSO) were proposed to enhance the discerning ability of the features and to select the discriminating features respectively. Three different emotional speech databases were utilized to gauge the proposed method. Extreme learning machine (ELM) was employed to classify the different types of emotions. Different experiments were conducted and the results show that the proposed method significantly improves the speech emotion recognition performance compared to previous works published in the literature. PMID:25799141

Identification of landscape features influencing gene flow: How useful are habitat selection models?

USGS Publications Warehouse

Roffler, Gretchen H.; Schwartz, Michael K.; Pilgrim, Kristy L.; Talbot, Sandra L.; Sage, Kevin; Adams, Layne G.; Luikart, Gordon

2016-01-01

Understanding how dispersal patterns are influenced by landscape heterogeneity is critical for modeling species connectivity. Resource selection function (RSF) models are increasingly used in landscape genetics approaches. However, because the ecological factors that drive habitat selection may be different from those influencing dispersal and gene flow, it is important to consider explicit assumptions and spatial scales of measurement. We calculated pairwise genetic distance among 301 Dall's sheep (Ovis dalli dalli) in southcentral Alaska using an intensive noninvasive sampling effort and 15 microsatellite loci. We used multiple regression of distance matrices to assess the correlation of pairwise genetic distance and landscape resistance derived from an RSF, and combinations of landscape features hypothesized to influence dispersal. Dall's sheep gene flow was positively correlated with steep slopes, moderate peak normalized difference vegetation indices (NDVI), and open land cover. Whereas RSF covariates were significant in predicting genetic distance, the RSF model itself was not significantly correlated with Dall's sheep gene flow, suggesting that certain habitat features important during summer (rugged terrain, mid-range elevation) were not influential to effective dispersal. This work underscores that consideration of both habitat selection and landscape genetics models may be useful in developing management strategies to both meet the immediate survival of a species and allow for long-term genetic connectivity.

Comprehensive comparative analysis and identification of RNA-binding protein domains: multi-class classification and feature selection.

PubMed

Jahandideh, Samad; Srinivasasainagendra, Vinodh; Zhi, Degui

2012-11-07

RNA-protein interaction plays an important role in various cellular processes, such as protein synthesis, gene regulation, post-transcriptional gene regulation, alternative splicing, and infections by RNA viruses. In this study, using Gene Ontology Annotated (GOA) and Structural Classification of Proteins (SCOP) databases an automatic procedure was designed to capture structurally solved RNA-binding protein domains in different subclasses. Subsequently, we applied tuned multi-class SVM (TMCSVM), Random Forest (RF), and multi-class ℓ1/ℓq-regularized logistic regression (MCRLR) for analysis and classifying RNA-binding protein domains based on a comprehensive set of sequence and structural features. In this study, we compared prediction accuracy of three different state-of-the-art predictor methods. From our results, TMCSVM outperforms the other methods and suggests the potential of TMCSVM as a useful tool for facilitating the multi-class prediction of RNA-binding protein domains. On the other hand, MCRLR by elucidating importance of features for their contribution in predictive accuracy of RNA-binding protein domains subclasses, helps us to provide some biological insights into the roles of sequences and structures in protein-RNA interactions.

Analyzing multicomponent receptive fields from neural responses to natural stimuli

PubMed Central

Rowekamp, Ryan; Sharpee, Tatyana O

2011-01-01

The challenge of building increasingly better models of neural responses to natural stimuli is to accurately estimate the multiple stimulus features that may jointly affect the neural spike probability. The selectivity for combinations of features is thought to be crucial for achieving classical properties of neural responses such as contrast invariance. The joint search for these multiple stimulus features is difficult because estimating spike probability as a multidimensional function of stimulus projections onto candidate relevant dimensions is subject to the curse of dimensionality. An attractive alternative is to search for relevant dimensions sequentially, as in projection pursuit regression. Here we demonstrate using analytic arguments and simulations of model cells that different types of sequential search strategies exhibit systematic biases when used with natural stimuli. Simulations show that joint optimization is feasible for up to three dimensions with current algorithms. When applied to the responses of V1 neurons to natural scenes, models based on three jointly optimized dimensions had better predictive power in a majority of cases compared to dimensions optimized sequentially, with different sequential methods yielding comparable results. Thus, although the curse of dimensionality remains, at least several relevant dimensions can be estimated by joint information maximization. PMID:21780916

Detection of chewing from piezoelectric film sensor signals using ensemble classifiers.

PubMed

Farooq, Muhammad; Sazonov, Edward

2016-08-01

Selection and use of pattern recognition algorithms is application dependent. In this work, we explored the use of several ensembles of weak classifiers to classify signals captured from a wearable sensor system to detect food intake based on chewing. Three sensor signals (Piezoelectric sensor, accelerometer, and hand to mouth gesture) were collected from 12 subjects in free-living conditions for 24 hrs. Sensor signals were divided into 10 seconds epochs and for each epoch combination of time and frequency domain features were computed. In this work, we present a comparison of three different ensemble techniques: boosting (AdaBoost), bootstrap aggregation (bagging) and stacking, each trained with 3 different weak classifiers (Decision Trees, Linear Discriminant Analysis (LDA) and Logistic Regression). Type of feature normalization used can also impact the classification results. For each ensemble method, three feature normalization techniques: (no-normalization, z-score normalization, and minmax normalization) were tested. A 12 fold cross-validation scheme was used to evaluate the performance of each model where the performance was evaluated in terms of precision, recall, and accuracy. Best results achieved here show an improvement of about 4% over our previous algorithms.

Improved Diagnostic Multimodal Biomarkers for Alzheimer's Disease and Mild Cognitive Impairment

PubMed Central

Martínez-Torteya, Antonio; Treviño, Víctor; Tamez-Peña, José G.

2015-01-01

The early diagnosis of Alzheimer's disease (AD) and mild cognitive impairment (MCI) is very important for treatment research and patient care purposes. Few biomarkers are currently considered in clinical settings, and their use is still optional. The objective of this work was to determine whether multimodal and nonpreviously AD associated features could improve the classification accuracy between AD, MCI, and healthy controls, which may impact future AD biomarkers. For this, Alzheimer's Disease Neuroimaging Initiative database was mined for case-control candidates. At least 652 baseline features extracted from MRI and PET analyses, biological samples, and clinical data up to February 2014 were used. A feature selection methodology that includes a genetic algorithm search coupled to a logistic regression classifier and forward and backward selection strategies was used to explore combinations of features. This generated diagnostic models with sizes ranging from 3 to 8, including well documented AD biomarkers, as well as unexplored image, biochemical, and clinical features. Accuracies of 0.85, 0.79, and 0.80 were achieved for HC-AD, HC-MCI, and MCI-AD classifications, respectively, when evaluated using a blind test set. In conclusion, a set of features provided additional and independent information to well-established AD biomarkers, aiding in the classification of MCI and AD. PMID:26106620

Complete hazard ranking to analyze right-censored data: An ALS survival study.

PubMed

Huang, Zhengnan; Zhang, Hongjiu; Boss, Jonathan; Goutman, Stephen A; Mukherjee, Bhramar; Dinov, Ivo D; Guan, Yuanfang

2017-12-01

Survival analysis represents an important outcome measure in clinical research and clinical trials; further, survival ranking may offer additional advantages in clinical trials. In this study, we developed GuanRank, a non-parametric ranking-based technique to transform patients' survival data into a linear space of hazard ranks. The transformation enables the utilization of machine learning base-learners including Gaussian process regression, Lasso, and random forest on survival data. The method was submitted to the DREAM Amyotrophic Lateral Sclerosis (ALS) Stratification Challenge. Ranked first place, the model gave more accurate ranking predictions on the PRO-ACT ALS dataset in comparison to Cox proportional hazard model. By utilizing right-censored data in its training process, the method demonstrated its state-of-the-art predictive power in ALS survival ranking. Its feature selection identified multiple important factors, some of which conflicts with previous studies.

Effective traffic features selection algorithm for cyber-attacks samples

NASA Astrophysics Data System (ADS)

Li, Yihong; Liu, Fangzheng; Du, Zhenyu

2018-05-01

By studying the defense scheme of Network attacks, this paper propose an effective traffic features selection algorithm based on k-means++ clustering to deal with the problem of high dimensionality of traffic features which extracted from cyber-attacks samples. Firstly, this algorithm divide the original feature set into attack traffic feature set and background traffic feature set by the clustering. Then, we calculates the variation of clustering performance after removing a certain feature. Finally, evaluating the degree of distinctiveness of the feature vector according to the result. Among them, the effective feature vector is whose degree of distinctiveness exceeds the set threshold. The purpose of this paper is to select out the effective features from the extracted original feature set. In this way, it can reduce the dimensionality of the features so as to reduce the space-time overhead of subsequent detection. The experimental results show that the proposed algorithm is feasible and it has some advantages over other selection algorithms.

A Study for the Feature Selection to Identify GIEMSA-Stained Human Chromosomes Based on Artificial Neural Network

DTIC Science & Technology

2001-10-25

neural network (ANN) has been adopted for the human chromosome classification. It is important to select optimum features for training neural network...Many studies for computer-based chromosome analysis have shown that it is possible to classify chromosomes into 24 subgroups. In addition, artificial

EXpectation Propagation LOgistic REgRession (EXPLORER): Distributed Privacy-Preserving Online Model Learning

PubMed Central

Wang, Shuang; Jiang, Xiaoqian; Wu, Yuan; Cui, Lijuan; Cheng, Samuel; Ohno-Machado, Lucila

2013-01-01

We developed an EXpectation Propagation LOgistic REgRession (EXPLORER) model for distributed privacy-preserving online learning. The proposed framework provides a high level guarantee for protecting sensitive information, since the information exchanged between the server and the client is the encrypted posterior distribution of coefficients. Through experimental results, EXPLORER shows the same performance (e.g., discrimination, calibration, feature selection etc.) as the traditional frequentist Logistic Regression model, but provides more flexibility in model updating. That is, EXPLORER can be updated one point at a time rather than having to retrain the entire data set when new observations are recorded. The proposed EXPLORER supports asynchronized communication, which relieves the participants from coordinating with one another, and prevents service breakdown from the absence of participants or interrupted communications. PMID:23562651

Feature Selection Methods for Zero-Shot Learning of Neural Activity.

PubMed

Caceres, Carlos A; Roos, Matthew J; Rupp, Kyle M; Milsap, Griffin; Crone, Nathan E; Wolmetz, Michael E; Ratto, Christopher R

2017-01-01

Dimensionality poses a serious challenge when making predictions from human neuroimaging data. Across imaging modalities, large pools of potential neural features (e.g., responses from particular voxels, electrodes, and temporal windows) have to be related to typically limited sets of stimuli and samples. In recent years, zero-shot prediction models have been introduced for mapping between neural signals and semantic attributes, which allows for classification of stimulus classes not explicitly included in the training set. While choices about feature selection can have a substantial impact when closed-set accuracy, open-set robustness, and runtime are competing design objectives, no systematic study of feature selection for these models has been reported. Instead, a relatively straightforward feature stability approach has been adopted and successfully applied across models and imaging modalities. To characterize the tradeoffs in feature selection for zero-shot learning, we compared correlation-based stability to several other feature selection techniques on comparable data sets from two distinct imaging modalities: functional Magnetic Resonance Imaging and Electrocorticography. While most of the feature selection methods resulted in similar zero-shot prediction accuracies and spatial/spectral patterns of selected features, there was one exception; A novel feature/attribute correlation approach was able to achieve those accuracies with far fewer features, suggesting the potential for simpler prediction models that yield high zero-shot classification accuracy.

Statistical performance of image cytometry for DNA, lipids, cytokeratin, & CD45 in a model system for circulation tumor cell detection.

PubMed

Futia, Gregory L; Schlaepfer, Isabel R; Qamar, Lubna; Behbakht, Kian; Gibson, Emily A

2017-07-01

Detection of circulating tumor cells (CTCs) in a blood sample is limited by the sensitivity and specificity of the biomarker panel used to identify CTCs over other blood cells. In this work, we present Bayesian theory that shows how test sensitivity and specificity set the rarity of cell that a test can detect. We perform our calculation of sensitivity and specificity on our image cytometry biomarker panel by testing on pure disease positive (D + ) populations (MCF7 cells) and pure disease negative populations (D - ) (leukocytes). In this system, we performed multi-channel confocal fluorescence microscopy to image biomarkers of DNA, lipids, CD45, and Cytokeratin. Using custom software, we segmented our confocal images into regions of interest consisting of individual cells and computed the image metrics of total signal, second spatial moment, spatial frequency second moment, and the product of the spatial-spatial frequency moments. We present our analysis of these 16 features. The best performing of the 16 features produced an average separation of three standard deviations between D + and D - and an average detectable rarity of ∼1 in 200. We performed multivariable regression and feature selection to combine multiple features for increased performance and showed an average separation of seven standard deviations between the D + and D - populations making our average detectable rarity of ∼1 in 480. Histograms and receiver operating characteristics (ROC) curves for these features and regressions are presented. We conclude that simple regression analysis holds promise to further improve the separation of rare cells in cytometry applications. © 2017 International Society for Advancement of Cytometry. © 2017 International Society for Advancement of Cytometry.

Selecting Feature Subsets Based on SVM-RFE and the Overlapping Ratio with Applications in Bioinformatics.

PubMed

Lin, Xiaohui; Li, Chao; Zhang, Yanhui; Su, Benzhe; Fan, Meng; Wei, Hai

2017-12-26

Feature selection is an important topic in bioinformatics. Defining informative features from complex high dimensional biological data is critical in disease study, drug development, etc. Support vector machine-recursive feature elimination (SVM-RFE) is an efficient feature selection technique that has shown its power in many applications. It ranks the features according to the recursive feature deletion sequence based on SVM. In this study, we propose a method, SVM-RFE-OA, which combines the classification accuracy rate and the average overlapping ratio of the samples to determine the number of features to be selected from the feature rank of SVM-RFE. Meanwhile, to measure the feature weights more accurately, we propose a modified SVM-RFE-OA (M-SVM-RFE-OA) algorithm that temporally screens out the samples lying in a heavy overlapping area in each iteration. The experiments on the eight public biological datasets show that the discriminative ability of the feature subset could be measured more accurately by combining the classification accuracy rate with the average overlapping degree of the samples compared with using the classification accuracy rate alone, and shielding the samples in the overlapping area made the calculation of the feature weights more stable and accurate. The methods proposed in this study can also be used with other RFE techniques to define potential biomarkers from big biological data.

Characterizing multivariate decoding models based on correlated EEG spectral features.

PubMed

McFarland, Dennis J

2013-07-01

Multivariate decoding methods are popular techniques for analysis of neurophysiological data. The present study explored potential interpretative problems with these techniques when predictors are correlated. Data from sensorimotor rhythm-based cursor control experiments was analyzed offline with linear univariate and multivariate models. Features were derived from autoregressive (AR) spectral analysis of varying model order which produced predictors that varied in their degree of correlation (i.e., multicollinearity). The use of multivariate regression models resulted in much better prediction of target position as compared to univariate regression models. However, with lower order AR features interpretation of the spectral patterns of the weights was difficult. This is likely to be due to the high degree of multicollinearity present with lower order AR features. Care should be exercised when interpreting the pattern of weights of multivariate models with correlated predictors. Comparison with univariate statistics is advisable. While multivariate decoding algorithms are very useful for prediction their utility for interpretation may be limited when predictors are correlated. Copyright © 2013 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

Classification of epileptic EEG signals based on simple random sampling and sequential feature selection.

PubMed

Ghayab, Hadi Ratham Al; Li, Yan; Abdulla, Shahab; Diykh, Mohammed; Wan, Xiangkui

2016-06-01

Electroencephalogram (EEG) signals are used broadly in the medical fields. The main applications of EEG signals are the diagnosis and treatment of diseases such as epilepsy, Alzheimer, sleep problems and so on. This paper presents a new method which extracts and selects features from multi-channel EEG signals. This research focuses on three main points. Firstly, simple random sampling (SRS) technique is used to extract features from the time domain of EEG signals. Secondly, the sequential feature selection (SFS) algorithm is applied to select the key features and to reduce the dimensionality of the data. Finally, the selected features are forwarded to a least square support vector machine (LS_SVM) classifier to classify the EEG signals. The LS_SVM classifier classified the features which are extracted and selected from the SRS and the SFS. The experimental results show that the method achieves 99.90, 99.80 and 100 % for classification accuracy, sensitivity and specificity, respectively.

Approximate median regression for complex survey data with skewed response.

PubMed

Fraser, Raphael André; Lipsitz, Stuart R; Sinha, Debajyoti; Fitzmaurice, Garrett M; Pan, Yi

2016-12-01

The ready availability of public-use data from various large national complex surveys has immense potential for the assessment of population characteristics using regression models. Complex surveys can be used to identify risk factors for important diseases such as cancer. Existing statistical methods based on estimating equations and/or utilizing resampling methods are often not valid with survey data due to complex survey design features. That is, stratification, multistage sampling, and weighting. In this article, we accommodate these design features in the analysis of highly skewed response variables arising from large complex surveys. Specifically, we propose a double-transform-both-sides (DTBS)'based estimating equations approach to estimate the median regression parameters of the highly skewed response; the DTBS approach applies the same Box-Cox type transformation twice to both the outcome and regression function. The usual sandwich variance estimate can be used in our approach, whereas a resampling approach would be needed for a pseudo-likelihood based on minimizing absolute deviations (MAD). Furthermore, the approach is relatively robust to the true underlying distribution, and has much smaller mean square error than a MAD approach. The method is motivated by an analysis of laboratory data on urinary iodine (UI) concentration from the National Health and Nutrition Examination Survey. © 2016, The International Biometric Society.

Approximate Median Regression for Complex Survey Data with Skewed Response

PubMed Central

Fraser, Raphael André; Lipsitz, Stuart R.; Sinha, Debajyoti; Fitzmaurice, Garrett M.; Pan, Yi

2016-01-01

Summary The ready availability of public-use data from various large national complex surveys has immense potential for the assessment of population characteristics using regression models. Complex surveys can be used to identify risk factors for important diseases such as cancer. Existing statistical methods based on estimating equations and/or utilizing resampling methods are often not valid with survey data due to complex survey design features. That is, stratification, multistage sampling and weighting. In this paper, we accommodate these design features in the analysis of highly skewed response variables arising from large complex surveys. Specifically, we propose a double-transform-both-sides (DTBS) based estimating equations approach to estimate the median regression parameters of the highly skewed response; the DTBS approach applies the same Box-Cox type transformation twice to both the outcome and regression function. The usual sandwich variance estimate can be used in our approach, whereas a resampling approach would be needed for a pseudo-likelihood based on minimizing absolute deviations (MAD). Furthermore, the approach is relatively robust to the true underlying distribution, and has much smaller mean square error than a MAD approach. The method is motivated by an analysis of laboratory data on urinary iodine (UI) concentration from the National Health and Nutrition Examination Survey. PMID:27062562

The Cross-Entropy Based Multi-Filter Ensemble Method for Gene Selection.

PubMed

Sun, Yingqiang; Lu, Chengbo; Li, Xiaobo

2018-05-17

The gene expression profile has the characteristics of a high dimension, low sample, and continuous type, and it is a great challenge to use gene expression profile data for the classification of tumor samples. This paper proposes a cross-entropy based multi-filter ensemble (CEMFE) method for microarray data classification. Firstly, multiple filters are used to select the microarray data in order to obtain a plurality of the pre-selected feature subsets with a different classification ability. The top N genes with the highest rank of each subset are integrated so as to form a new data set. Secondly, the cross-entropy algorithm is used to remove the redundant data in the data set. Finally, the wrapper method, which is based on forward feature selection, is used to select the best feature subset. The experimental results show that the proposed method is more efficient than other gene selection methods and that it can achieve a higher classification accuracy under fewer characteristic genes.

Exploring Sampling in the Detection of Multicategory EEG Signals

PubMed Central

Siuly, Siuly; Kabir, Enamul; Wang, Hua; Zhang, Yanchun

2015-01-01

The paper presents a structure based on samplings and machine leaning techniques for the detection of multicategory EEG signals where random sampling (RS) and optimal allocation sampling (OS) are explored. In the proposed framework, before using the RS and OS scheme, the entire EEG signals of each class are partitioned into several groups based on a particular time period. The RS and OS schemes are used in order to have representative observations from each group of each category of EEG data. Then all of the selected samples by the RS from the groups of each category are combined in a one set named RS set. In the similar way, for the OS scheme, an OS set is obtained. Then eleven statistical features are extracted from the RS and OS set, separately. Finally this study employs three well-known classifiers: k-nearest neighbor (k-NN), multinomial logistic regression with a ridge estimator (MLR), and support vector machine (SVM) to evaluate the performance for the RS and OS feature set. The experimental outcomes demonstrate that the RS scheme well represents the EEG signals and the k-NN with the RS is the optimum choice for detection of multicategory EEG signals. PMID:25977705

RBF kernel based support vector regression to estimate the blood volume and heart rate responses during hemodialysis.

PubMed

Javed, Faizan; Chan, Gregory S H; Savkin, Andrey V; Middleton, Paul M; Malouf, Philip; Steel, Elizabeth; Mackie, James; Lovell, Nigel H

2009-01-01

This paper uses non-linear support vector regression (SVR) to model the blood volume and heart rate (HR) responses in 9 hemodynamically stable kidney failure patients during hemodialysis. Using radial bias function (RBF) kernels the non-parametric models of relative blood volume (RBV) change with time as well as percentage change in HR with respect to RBV were obtained. The e-insensitivity based loss function was used for SVR modeling. Selection of the design parameters which includes capacity (C), insensitivity region (e) and the RBF kernel parameter (sigma) was made based on a grid search approach and the selected models were cross-validated using the average mean square error (AMSE) calculated from testing data based on a k-fold cross-validation technique. Linear regression was also applied to fit the curves and the AMSE was calculated for comparison with SVR. For the model based on RBV with time, SVR gave a lower AMSE for both training (AMSE=1.5) as well as testing data (AMSE=1.4) compared to linear regression (AMSE=1.8 and 1.5). SVR also provided a better fit for HR with RBV for both training as well as testing data (AMSE=15.8 and 16.4) compared to linear regression (AMSE=25.2 and 20.1).

Selecting relevant 3D image features of margin sharpness and texture for lung nodule retrieval.

PubMed

Ferreira, José Raniery; de Azevedo-Marques, Paulo Mazzoncini; Oliveira, Marcelo Costa

2017-03-01

Lung cancer is the leading cause of cancer-related deaths in the world. Its diagnosis is a challenge task to specialists due to several aspects on the classification of lung nodules. Therefore, it is important to integrate content-based image retrieval methods on the lung nodule classification process, since they are capable of retrieving similar cases from databases that were previously diagnosed. However, this mechanism depends on extracting relevant image features in order to obtain high efficiency. The goal of this paper is to perform the selection of 3D image features of margin sharpness and texture that can be relevant on the retrieval of similar cancerous and benign lung nodules. A total of 48 3D image attributes were extracted from the nodule volume. Border sharpness features were extracted from perpendicular lines drawn over the lesion boundary. Second-order texture features were extracted from a cooccurrence matrix. Relevant features were selected by a correlation-based method and a statistical significance analysis. Retrieval performance was assessed according to the nodule's potential malignancy on the 10 most similar cases and by the parameters of precision and recall. Statistical significant features reduced retrieval performance. Correlation-based method selected 2 margin sharpness attributes and 6 texture attributes and obtained higher precision compared to all 48 extracted features on similar nodule retrieval. Feature space dimensionality reduction of 83 % obtained higher retrieval performance and presented to be a computationaly low cost method of retrieving similar nodules for the diagnosis of lung cancer.

Postharvest monitoring of organic potato (cv. Anuschka) during hot-air drying using visible-NIR hyperspectral imaging.

PubMed

Moscetti, Roberto; Sturm, Barbara; Crichton, Stuart Oj; Amjad, Waseem; Massantini, Riccardo

2018-05-01

The potential of hyperspectral imaging (500-1010 nm) was evaluated for monitoring of the quality of potato slices (var. Anuschka) of 5, 7 and 9 mm thickness subjected to air drying at 50 °C. The study investigated three different feature selection methods for the prediction of dry basis moisture content and colour of potato slices using partial least squares regression (PLS). The feature selection strategies tested include interval PLS regression (iPLS), and differences and ratios between raw reflectance values for each possible pair of wavelengths (R[λ 1 ]-R[λ 2 ] and R[λ 1 ]:R[λ 2 ], respectively). Moreover, the combination of spectral and spatial domains was tested. Excellent results were obtained using the iPLS algorithm. However, features from both datasets of raw reflectance differences and ratios represent suitable alternatives for development of low-complex prediction models. Finally, the dry basis moisture content was high accurately predicted by combining spectral data (i.e. R[511 nm]-R[994 nm]) and spatial domain (i.e. relative area shrinkage of slice). Modelling the data acquired during drying through hyperspectral imaging can provide useful information concerning the chemical and physicochemical changes of the product. With all this information, the proposed approach lays the foundations for a more efficient smart dryer that can be designed and its process optimized for drying of potato slices. © 2017 Society of Chemical Industry. © 2017 Society of Chemical Industry.

Effective and extensible feature extraction method using genetic algorithm-based frequency-domain feature search for epileptic EEG multiclassification

PubMed Central

Wen, Tingxi; Zhang, Zhongnan

2017-01-01

Abstract In this paper, genetic algorithm-based frequency-domain feature search (GAFDS) method is proposed for the electroencephalogram (EEG) analysis of epilepsy. In this method, frequency-domain features are first searched and then combined with nonlinear features. Subsequently, these features are selected and optimized to classify EEG signals. The extracted features are analyzed experimentally. The features extracted by GAFDS show remarkable independence, and they are superior to the nonlinear features in terms of the ratio of interclass distance and intraclass distance. Moreover, the proposed feature search method can search for features of instantaneous frequency in a signal after Hilbert transformation. The classification results achieved using these features are reasonable; thus, GAFDS exhibits good extensibility. Multiple classical classifiers (i.e., k-nearest neighbor, linear discriminant analysis, decision tree, AdaBoost, multilayer perceptron, and Naïve Bayes) achieve satisfactory classification accuracies by using the features generated by the GAFDS method and the optimized feature selection. The accuracies for 2-classification and 3-classification problems may reach up to 99% and 97%, respectively. Results of several cross-validation experiments illustrate that GAFDS is effective in the extraction of effective features for EEG classification. Therefore, the proposed feature selection and optimization model can improve classification accuracy. PMID:28489789

Effective and extensible feature extraction method using genetic algorithm-based frequency-domain feature search for epileptic EEG multiclassification.

PubMed

Wen, Tingxi; Zhang, Zhongnan

2017-05-01

In this paper, genetic algorithm-based frequency-domain feature search (GAFDS) method is proposed for the electroencephalogram (EEG) analysis of epilepsy. In this method, frequency-domain features are first searched and then combined with nonlinear features. Subsequently, these features are selected and optimized to classify EEG signals. The extracted features are analyzed experimentally. The features extracted by GAFDS show remarkable independence, and they are superior to the nonlinear features in terms of the ratio of interclass distance and intraclass distance. Moreover, the proposed feature search method can search for features of instantaneous frequency in a signal after Hilbert transformation. The classification results achieved using these features are reasonable; thus, GAFDS exhibits good extensibility. Multiple classical classifiers (i.e., k-nearest neighbor, linear discriminant analysis, decision tree, AdaBoost, multilayer perceptron, and Naïve Bayes) achieve satisfactory classification accuracies by using the features generated by the GAFDS method and the optimized feature selection. The accuracies for 2-classification and 3-classification problems may reach up to 99% and 97%, respectively. Results of several cross-validation experiments illustrate that GAFDS is effective in the extraction of effective features for EEG classification. Therefore, the proposed feature selection and optimization model can improve classification accuracy.

Decreased activation and subsyndromal manic symptoms predict lower remission rates in bipolar depression.

PubMed

Caldieraro, Marco Antonio; Walsh, Samantha; Deckersbach, Thilo; Bobo, William V; Gao, Keming; Ketter, Terence A; Shelton, Richard C; Reilly-Harrington, Noreen A; Tohen, Mauricio; Calabrese, Joseph R; Thase, Michael E; Kocsis, James H; Sylvia, Louisa G; Nierenberg, Andrew A

2017-11-01

Activation encompasses energy and activity and is a central feature of bipolar disorder. However, the impact of activation on treatment response of bipolar depression requires further exploration. The aims of this study were to assess the association of decreased activation and sustained remission in bipolar depression and test for factors that could affect this association. We assessed participants with Diagnostic and Statistical Manual of Mental Disorders (4th ed) bipolar depression ( n = 303) included in a comparative effectiveness study of lithium- and quetiapine-based treatments (the Bipolar CHOICE study). Activation was evaluated using items from the Bipolar Inventory of Symptoms Scale. The selection of these items was based on a dimension of energy and interest symptoms associated with poorer treatment response in major depression. Decreased activation was associated with lower remission rates in the raw analyses and in a logistic regression model adjusted for baseline severity and subsyndromal manic symptoms (odds ratio = 0.899; p = 0.015). The manic features also predicted lower remission (odds ratio = 0.934; p < 0.001). Remission rates were similar in the two treatment groups. Decreased activation and subsyndromal manic symptoms predict lower remission rates in bipolar depression. Patients with these features may require specific treatment approaches, but new studies are necessary to identify treatments that could improve outcomes in this population.

The value of nodal information in predicting lung cancer relapse using 4DPET/4DCT

DOE Office of Scientific and Technical Information (OSTI.GOV)

Li, Heyse, E-mail: heyse.li@mail.utoronto.ca; Becker, Nathan; Raman, Srinivas

2015-08-15

Purpose: There is evidence that computed tomography (CT) and positron emission tomography (PET) imaging metrics are prognostic and predictive in nonsmall cell lung cancer (NSCLC) treatment outcomes. However, few studies have explored the use of standardized uptake value (SUV)-based image features of nodal regions as predictive features. The authors investigated and compared the use of tumor and node image features extracted from the radiotherapy target volumes to predict relapse in a cohort of NSCLC patients undergoing chemoradiation treatment. Methods: A prospective cohort of 25 patients with locally advanced NSCLC underwent 4DPET/4DCT imaging for radiation planning. Thirty-seven image features were derivedmore » from the CT-defined volumes and SUVs of the PET image from both the tumor and nodal target regions. The machine learning methods of logistic regression and repeated stratified five-fold cross-validation (CV) were used to predict local and overall relapses in 2 yr. The authors used well-known feature selection methods (Spearman’s rank correlation, recursive feature elimination) within each fold of CV. Classifiers were ranked on their Matthew’s correlation coefficient (MCC) after CV. Area under the curve, sensitivity, and specificity values are also presented. Results: For predicting local relapse, the best classifier found had a mean MCC of 0.07 and was composed of eight tumor features. For predicting overall relapse, the best classifier found had a mean MCC of 0.29 and was composed of a single feature: the volume greater than 0.5 times the maximum SUV (N). Conclusions: The best classifier for predicting local relapse had only tumor features. In contrast, the best classifier for predicting overall relapse included a node feature. Overall, the methods showed that nodes add value in predicting overall relapse but not local relapse.« less

Feature-Selective Attention Adaptively Shifts Noise Correlations in Primary Auditory Cortex.

PubMed

Downer, Joshua D; Rapone, Brittany; Verhein, Jessica; O'Connor, Kevin N; Sutter, Mitchell L

2017-05-24

Sensory environments often contain an overwhelming amount of information, with both relevant and irrelevant information competing for neural resources. Feature attention mediates this competition by selecting the sensory features needed to form a coherent percept. How attention affects the activity of populations of neurons to support this process is poorly understood because population coding is typically studied through simulations in which one sensory feature is encoded without competition. Therefore, to study the effects of feature attention on population-based neural coding, investigations must be extended to include stimuli with both relevant and irrelevant features. We measured noise correlations ( r noise ) within small neural populations in primary auditory cortex while rhesus macaques performed a novel feature-selective attention task. We found that the effect of feature-selective attention on r noise depended not only on the population tuning to the attended feature, but also on the tuning to the distractor feature. To attempt to explain how these observed effects might support enhanced perceptual performance, we propose an extension of a simple and influential model in which shifts in r noise can simultaneously enhance the representation of the attended feature while suppressing the distractor. These findings present a novel mechanism by which attention modulates neural populations to support sensory processing in cluttered environments. SIGNIFICANCE STATEMENT Although feature-selective attention constitutes one of the building blocks of listening in natural environments, its neural bases remain obscure. To address this, we developed a novel auditory feature-selective attention task and measured noise correlations ( r noise ) in rhesus macaque A1 during task performance. Unlike previous studies showing that the effect of attention on r noise depends on population tuning to the attended feature, we show that the effect of attention depends on the tuning to the distractor feature as well. We suggest that these effects represent an efficient process by which sensory cortex simultaneously enhances relevant information and suppresses irrelevant information. Copyright © 2017 the authors 0270-6474/17/375378-15$15.00/0.

Feature-Selective Attention Adaptively Shifts Noise Correlations in Primary Auditory Cortex

PubMed Central

2017-01-01

Sensory environments often contain an overwhelming amount of information, with both relevant and irrelevant information competing for neural resources. Feature attention mediates this competition by selecting the sensory features needed to form a coherent percept. How attention affects the activity of populations of neurons to support this process is poorly understood because population coding is typically studied through simulations in which one sensory feature is encoded without competition. Therefore, to study the effects of feature attention on population-based neural coding, investigations must be extended to include stimuli with both relevant and irrelevant features. We measured noise correlations (rnoise) within small neural populations in primary auditory cortex while rhesus macaques performed a novel feature-selective attention task. We found that the effect of feature-selective attention on rnoise depended not only on the population tuning to the attended feature, but also on the tuning to the distractor feature. To attempt to explain how these observed effects might support enhanced perceptual performance, we propose an extension of a simple and influential model in which shifts in rnoise can simultaneously enhance the representation of the attended feature while suppressing the distractor. These findings present a novel mechanism by which attention modulates neural populations to support sensory processing in cluttered environments. SIGNIFICANCE STATEMENT Although feature-selective attention constitutes one of the building blocks of listening in natural environments, its neural bases remain obscure. To address this, we developed a novel auditory feature-selective attention task and measured noise correlations (rnoise) in rhesus macaque A1 during task performance. Unlike previous studies showing that the effect of attention on rnoise depends on population tuning to the attended feature, we show that the effect of attention depends on the tuning to the distractor feature as well. We suggest that these effects represent an efficient process by which sensory cortex simultaneously enhances relevant information and suppresses irrelevant information. PMID:28432139

Classification of early-stage non-small cell lung cancer by weighing gene expression profiles with connectivity information.

PubMed

Zhang, Ao; Tian, Suyan

2018-05-01

Pathway-based feature selection algorithms, which utilize biological information contained in pathways to guide which features/genes should be selected, have evolved quickly and become widespread in the field of bioinformatics. Based on how the pathway information is incorporated, we classify pathway-based feature selection algorithms into three major categories-penalty, stepwise forward, and weighting. Compared to the first two categories, the weighting methods have been underutilized even though they are usually the simplest ones. In this article, we constructed three different genes' connectivity information-based weights for each gene and then conducted feature selection upon the resulting weighted gene expression profiles. Using both simulations and a real-world application, we have demonstrated that when the data-driven connectivity information constructed from the data of specific disease under study is considered, the resulting weighted gene expression profiles slightly outperform the original expression profiles. In summary, a big challenge faced by the weighting method is how to estimate pathway knowledge-based weights more accurately and precisely. Only until the issue is conquered successfully will wide utilization of the weighting methods be impossible. © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

Greedy feature selection for glycan chromatography data with the generalized Dirichlet distribution

PubMed Central

2013-01-01

Background Glycoproteins are involved in a diverse range of biochemical and biological processes. Changes in protein glycosylation are believed to occur in many diseases, particularly during cancer initiation and progression. The identification of biomarkers for human disease states is becoming increasingly important, as early detection is key to improving survival and recovery rates. To this end, the serum glycome has been proposed as a potential source of biomarkers for different types of cancers. High-throughput hydrophilic interaction liquid chromatography (HILIC) technology for glycan analysis allows for the detailed quantification of the glycan content in human serum. However, the experimental data from this analysis is compositional by nature. Compositional data are subject to a constant-sum constraint, which restricts the sample space to a simplex. Statistical analysis of glycan chromatography datasets should account for their unusual mathematical properties. As the volume of glycan HILIC data being produced increases, there is a considerable need for a framework to support appropriate statistical analysis. Proposed here is a methodology for feature selection in compositional data. The principal objective is to provide a template for the analysis of glycan chromatography data that may be used to identify potential glycan biomarkers. Results A greedy search algorithm, based on the generalized Dirichlet distribution, is carried out over the feature space to search for the set of “grouping variables” that best discriminate between known group structures in the data, modelling the compositional variables using beta distributions. The algorithm is applied to two glycan chromatography datasets. Statistical classification methods are used to test the ability of the selected features to differentiate between known groups in the data. Two well-known methods are used for comparison: correlation-based feature selection (CFS) and recursive partitioning (rpart). CFS is a feature selection method, while recursive partitioning is a learning tree algorithm that has been used for feature selection in the past. Conclusions The proposed feature selection method performs well for both glycan chromatography datasets. It is computationally slower, but results in a lower misclassification rate and a higher sensitivity rate than both correlation-based feature selection and the classification tree method. PMID:23651459

Examining the relation between rock mass cuttability index and rock drilling properties

NASA Astrophysics Data System (ADS)

Yetkin, Mustafa E.; Özfırat, M. Kemal; Yenice, Hayati; Şimşir, Ferhan; Kahraman, Bayram

2016-12-01

Drilling rate is a substantial index value in drilling and excavation operations at mining. It is not only a help in determining physical and mechanical features of rocks, but also delivers strong estimations about instantaneous cutting rates. By this way, work durations to be finished on time, proper machine/equipment selection and efficient excavation works can be achieved. In this study, physical and mechanical properties of surrounding rocks and ore zones are determined by investigations carried out on specimens taken from an underground ore mine. Later, relationships among rock mass classifications, drillability rates, cuttability, and abrasivity have been investigated using multi regression analysis. As a result, equations having high regression rates have been found out among instantaneous cutting rates and geomechanical properties of rocks. Moreover, excavation machine selection for the study area has been made at the best possible interval.

Use of probabilistic weights to enhance linear regression myoelectric control

NASA Astrophysics Data System (ADS)

Smith, Lauren H.; Kuiken, Todd A.; Hargrove, Levi J.

2015-12-01

Objective. Clinically available prostheses for transradial amputees do not allow simultaneous myoelectric control of degrees of freedom (DOFs). Linear regression methods can provide simultaneous myoelectric control, but frequently also result in difficulty with isolating individual DOFs when desired. This study evaluated the potential of using probabilistic estimates of categories of gross prosthesis movement, which are commonly used in classification-based myoelectric control, to enhance linear regression myoelectric control. Approach. Gaussian models were fit to electromyogram (EMG) feature distributions for three movement classes at each DOF (no movement, or movement in either direction) and used to weight the output of linear regression models by the probability that the user intended the movement. Eight able-bodied and two transradial amputee subjects worked in a virtual Fitts’ law task to evaluate differences in controllability between linear regression and probability-weighted regression for an intramuscular EMG-based three-DOF wrist and hand system. Main results. Real-time and offline analyses in able-bodied subjects demonstrated that probability weighting improved performance during single-DOF tasks (p < 0.05) by preventing extraneous movement at additional DOFs. Similar results were seen in experiments with two transradial amputees. Though goodness-of-fit evaluations suggested that the EMG feature distributions showed some deviations from the Gaussian, equal-covariance assumptions used in this experiment, the assumptions were sufficiently met to provide improved performance compared to linear regression control. Significance. Use of probability weights can improve the ability to isolate individual during linear regression myoelectric control, while maintaining the ability to simultaneously control multiple DOFs.

Improving lung cancer prognosis assessment by incorporating synthetic minority oversampling technique and score fusion method

DOE Office of Scientific and Technical Information (OSTI.GOV)

Yan, Shiju; Qian, Wei; Guan, Yubao

2016-06-15

Purpose: This study aims to investigate the potential to improve lung cancer recurrence risk prediction performance for stage I NSCLS patients by integrating oversampling, feature selection, and score fusion techniques and develop an optimal prediction model. Methods: A dataset involving 94 early stage lung cancer patients was retrospectively assembled, which includes CT images, nine clinical and biological (CB) markers, and outcome of 3-yr disease-free survival (DFS) after surgery. Among the 94 patients, 74 remained DFS and 20 had cancer recurrence. Applying a computer-aided detection scheme, tumors were segmented from the CT images and 35 quantitative image (QI) features were initiallymore » computed. Two normalized Gaussian radial basis function network (RBFN) based classifiers were built based on QI features and CB markers separately. To improve prediction performance, the authors applied a synthetic minority oversampling technique (SMOTE) and a BestFirst based feature selection method to optimize the classifiers and also tested fusion methods to combine QI and CB based prediction results. Results: Using a leave-one-case-out cross-validation (K-fold cross-validation) method, the computed areas under a receiver operating characteristic curve (AUCs) were 0.716 ± 0.071 and 0.642 ± 0.061, when using the QI and CB based classifiers, respectively. By fusion of the scores generated by the two classifiers, AUC significantly increased to 0.859 ± 0.052 (p < 0.05) with an overall prediction accuracy of 89.4%. Conclusions: This study demonstrated the feasibility of improving prediction performance by integrating SMOTE, feature selection, and score fusion techniques. Combining QI features and CB markers and performing SMOTE prior to feature selection in classifier training enabled RBFN based classifier to yield improved prediction accuracy.« less

Feature Selection based on Machine Learning in MRIs for Hippocampal Segmentation

NASA Astrophysics Data System (ADS)

Tangaro, Sabina; Amoroso, Nicola; Brescia, Massimo; Cavuoti, Stefano; Chincarini, Andrea; Errico, Rosangela; Paolo, Inglese; Longo, Giuseppe; Maglietta, Rosalia; Tateo, Andrea; Riccio, Giuseppe; Bellotti, Roberto

2015-01-01

Neurodegenerative diseases are frequently associated with structural changes in the brain. Magnetic resonance imaging (MRI) scans can show these variations and therefore can be used as a supportive feature for a number of neurodegenerative diseases. The hippocampus has been known to be a biomarker for Alzheimer disease and other neurological and psychiatric diseases. However, it requires accurate, robust, and reproducible delineation of hippocampal structures. Fully automatic methods are usually the voxel based approach; for each voxel a number of local features were calculated. In this paper, we compared four different techniques for feature selection from a set of 315 features extracted for each voxel: (i) filter method based on the Kolmogorov-Smirnov test; two wrapper methods, respectively, (ii) sequential forward selection and (iii) sequential backward elimination; and (iv) embedded method based on the Random Forest Classifier on a set of 10 T1-weighted brain MRIs and tested on an independent set of 25 subjects. The resulting segmentations were compared with manual reference labelling. By using only 23 feature for each voxel (sequential backward elimination) we obtained comparable state-of-the-art performances with respect to the standard tool FreeSurfer.

Hybrid feature selection for supporting lightweight intrusion detection systems

NASA Astrophysics Data System (ADS)

Song, Jianglong; Zhao, Wentao; Liu, Qiang; Wang, Xin

2017-08-01

Redundant and irrelevant features not only cause high resource consumption but also degrade the performance of Intrusion Detection Systems (IDS), especially when coping with big data. These features slow down the process of training and testing in network traffic classification. Therefore, a hybrid feature selection approach in combination with wrapper and filter selection is designed in this paper to build a lightweight intrusion detection system. Two main phases are involved in this method. The first phase conducts a preliminary search for an optimal subset of features, in which the chi-square feature selection is utilized. The selected set of features from the previous phase is further refined in the second phase in a wrapper manner, in which the Random Forest(RF) is used to guide the selection process and retain an optimized set of features. After that, we build an RF-based detection model and make a fair comparison with other approaches. The experimental results on NSL-KDD datasets show that our approach results are in higher detection accuracy as well as faster training and testing processes.

Feature Selection for Classification of Polar Regions Using a Fuzzy Expert System

NASA Technical Reports Server (NTRS)

Penaloza, Mauel A.; Welch, Ronald M.

1996-01-01

Labeling, feature selection, and the choice of classifier are critical elements for classification of scenes and for image understanding. This study examines several methods for feature selection in polar regions, including the list, of a fuzzy logic-based expert system for further refinement of a set of selected features. Six Advanced Very High Resolution Radiometer (AVHRR) Local Area Coverage (LAC) arctic scenes are classified into nine classes: water, snow / ice, ice cloud, land, thin stratus, stratus over water, cumulus over water, textured snow over water, and snow-covered mountains. Sixty-seven spectral and textural features are computed and analyzed by the feature selection algorithms. The divergence, histogram analysis, and discriminant analysis approaches are intercompared for their effectiveness in feature selection. The fuzzy expert system method is used not only to determine the effectiveness of each approach in classifying polar scenes, but also to further reduce the features into a more optimal set. For each selection method,features are ranked from best to worst, and the best half of the features are selected. Then, rules using these selected features are defined. The results of running the fuzzy expert system with these rules show that the divergence method produces the best set features, not only does it produce the highest classification accuracy, but also it has the lowest computation requirements. A reduction of the set of features produced by the divergence method using the fuzzy expert system results in an overall classification accuracy of over 95 %. However, this increase of accuracy has a high computation cost.

An effective biometric discretization approach to extract highly discriminative, informative, and privacy-protective binary representation

NASA Astrophysics Data System (ADS)

Lim, Meng-Hui; Teoh, Andrew Beng Jin

2011-12-01

Biometric discretization derives a binary string for each user based on an ordered set of biometric features. This representative string ought to be discriminative, informative, and privacy protective when it is employed as a cryptographic key in various security applications upon error correction. However, it is commonly believed that satisfying the first and the second criteria simultaneously is not feasible, and a tradeoff between them is always definite. In this article, we propose an effective fixed bit allocation-based discretization approach which involves discriminative feature extraction, discriminative feature selection, unsupervised quantization (quantization that does not utilize class information), and linearly separable subcode (LSSC)-based encoding to fulfill all the ideal properties of a binary representation extracted for cryptographic applications. In addition, we examine a number of discriminative feature-selection measures for discretization and identify the proper way of setting an important feature-selection parameter. Encouraging experimental results vindicate the feasibility of our approach.

Feature Selection for Object-Based Classification of High-Resolution Remote Sensing Images Based on the Combination of a Genetic Algorithm and Tabu Search

PubMed Central

Shi, Lei; Wan, Youchuan; Gao, Xianjun

2018-01-01

In object-based image analysis of high-resolution images, the number of features can reach hundreds, so it is necessary to perform feature reduction prior to classification. In this paper, a feature selection method based on the combination of a genetic algorithm (GA) and tabu search (TS) is presented. The proposed GATS method aims to reduce the premature convergence of the GA by the use of TS. A prematurity index is first defined to judge the convergence situation during the search. When premature convergence does take place, an improved mutation operator is executed, in which TS is performed on individuals with higher fitness values. As for the other individuals with lower fitness values, mutation with a higher probability is carried out. Experiments using the proposed GATS feature selection method and three other methods, a standard GA, the multistart TS method, and ReliefF, were conducted on WorldView-2 and QuickBird images. The experimental results showed that the proposed method outperforms the other methods in terms of the final classification accuracy. PMID:29581721

A Feature Selection Method Based on Fisher's Discriminant Ratio for Text Sentiment Classification

NASA Astrophysics Data System (ADS)

Wang, Suge; Li, Deyu; Wei, Yingjie; Li, Hongxia

With the rapid growth of e-commerce, product reviews on the Web have become an important information source for customers' decision making when they intend to buy some product. As the reviews are often too many for customers to go through, how to automatically classify them into different sentiment orientation categories (i.e. positive/negative) has become a research problem. In this paper, based on Fisher's discriminant ratio, an effective feature selection method is proposed for product review text sentiment classification. In order to validate the validity of the proposed method, we compared it with other methods respectively based on information gain and mutual information while support vector machine is adopted as the classifier. In this paper, 6 subexperiments are conducted by combining different feature selection methods with 2 kinds of candidate feature sets. Under 1006 review documents of cars, the experimental results indicate that the Fisher's discriminant ratio based on word frequency estimation has the best performance with F value 83.3% while the candidate features are the words which appear in both positive and negative texts.

Face aging effect simulation model based on multilayer representation and shearlet transform

NASA Astrophysics Data System (ADS)

Li, Yuancheng; Li, Yan

2017-09-01

In order to extract detailed facial features, we build a face aging effect simulation model based on multilayer representation and shearlet transform. The face is divided into three layers: the global layer of the face, the local features layer, and texture layer, which separately establishes the aging model. First, the training samples are classified according to different age groups, and we use active appearance model (AAM) at the global level to obtain facial features. The regression equations of shape and texture with age are obtained by fitting the support vector machine regression, which is based on the radial basis function. We use AAM to simulate the aging of facial organs. Then, for the texture detail layer, we acquire the significant high-frequency characteristic components of the face by using the multiscale shearlet transform. Finally, we get the last simulated aging images of the human face by the fusion algorithm. Experiments are carried out on the FG-NET dataset, and the experimental results show that the simulated face images have less differences from the original image and have a good face aging simulation effect.

Aboveground Biomass Estimation Using Reconstructed Feature of Airborne Discrete-Return LIDAR by Auto-Encoder Neural Network

NASA Astrophysics Data System (ADS)

Li, T.; Wang, Z.; Peng, J.

2018-04-01

Aboveground biomass (AGB) estimation is critical for quantifying carbon stocks and essential for evaluating carbon cycle. In recent years, airborne LiDAR shows its great ability for highly-precision AGB estimation. Most of the researches estimate AGB by the feature metrics extracted from the canopy height distribution of the point cloud which calculated based on precise digital terrain model (DTM). However, if forest canopy density is high, the probability of the LiDAR signal penetrating the canopy is lower, resulting in ground points is not enough to establish DTM. Then the distribution of forest canopy height is imprecise and some critical feature metrics which have a strong correlation with biomass such as percentiles, maximums, means and standard deviations of canopy point cloud can hardly be extracted correctly. In order to address this issue, we propose a strategy of first reconstructing LiDAR feature metrics through Auto-Encoder neural network and then using the reconstructed feature metrics to estimate AGB. To assess the prediction ability of the reconstructed feature metrics, both original and reconstructed feature metrics were regressed against field-observed AGB using the multiple stepwise regression (MS) and the partial least squares regression (PLS) respectively. The results showed that the estimation model using reconstructed feature metrics improved R2 by 5.44 %, 18.09 %, decreased RMSE value by 10.06 %, 22.13 % and reduced RMSEcv by 10.00 %, 21.70 % for AGB, respectively. Therefore, reconstructing LiDAR point feature metrics has potential for addressing AGB estimation challenge in dense canopy area.

Parenchymal texture analysis in digital mammography: A fully automated pipeline for breast cancer risk assessment.

PubMed

Zheng, Yuanjie; Keller, Brad M; Ray, Shonket; Wang, Yan; Conant, Emily F; Gee, James C; Kontos, Despina

2015-07-01

Mammographic percent density (PD%) is known to be a strong risk factor for breast cancer. Recent studies also suggest that parenchymal texture features, which are more granular descriptors of the parenchymal pattern, can provide additional information about breast cancer risk. To date, most studies have measured mammographic texture within selected regions of interest (ROIs) in the breast, which cannot adequately capture the complexity of the parenchymal pattern throughout the whole breast. To better characterize patterns of the parenchymal tissue, the authors have developed a fully automated software pipeline based on a novel lattice-based strategy to extract a range of parenchymal texture features from the entire breast region. Digital mammograms from 106 cases with 318 age-matched controls were retrospectively analyzed. The lattice-based approach is based on a regular grid virtually overlaid on each mammographic image. Texture features are computed from the intersection (i.e., lattice) points of the grid lines within the breast, using a local window centered at each lattice point. Using this strategy, a range of statistical (gray-level histogram, co-occurrence, and run-length) and structural (edge-enhancing, local binary pattern, and fractal dimension) features are extracted. To cover the entire breast, the size of the local window for feature extraction is set equal to the lattice grid spacing and optimized experimentally by evaluating different windows sizes. The association between their lattice-based texture features and breast cancer was evaluated using logistic regression with leave-one-out cross validation and further compared to that of breast PD% and commonly used single-ROI texture features extracted from the retroareolar or the central breast region. Classification performance was evaluated using the area under the curve (AUC) of the receiver operating characteristic (ROC). DeLong's test was used to compare the different ROCs in terms of AUC performance. The average univariate performance of the lattice-based features is higher when extracted from smaller than larger window sizes. While not every individual texture feature is superior to breast PD% (AUC: 0.59, STD: 0.03), their combination in multivariate analysis has significantly better performance (AUC: 0.85, STD: 0.02, p < 0.001). The lattice-based texture features also outperform the single-ROI texture features when extracted from the retroareolar or the central breast region (AUC: 0.60-0.74, STD: 0.03). Adding breast PD% does not make a significant performance improvement to the lattice-based texture features or the single-ROI features (p > 0.05). The proposed lattice-based strategy for mammographic texture analysis enables to characterize the parenchymal pattern over the entire breast. As such, these features provide richer information compared to currently used descriptors and may ultimately improve breast cancer risk assessment. Larger studies are warranted to validate these findings and also compare to standard demographic and reproductive risk factors.

Reducing Sweeping Frequencies in Microwave NDT Employing Machine Learning Feature Selection

PubMed Central

Moomen, Abdelniser; Ali, Abdulbaset; Ramahi, Omar M.

2016-01-01

Nondestructive Testing (NDT) assessment of materials’ health condition is useful for classifying healthy from unhealthy structures or detecting flaws in metallic or dielectric structures. Performing structural health testing for coated/uncoated metallic or dielectric materials with the same testing equipment requires a testing method that can work on metallics and dielectrics such as microwave testing. Reducing complexity and expenses associated with current diagnostic practices of microwave NDT of structural health requires an effective and intelligent approach based on feature selection and classification techniques of machine learning. Current microwave NDT methods in general based on measuring variation in the S-matrix over the entire operating frequency ranges of the sensors. For instance, assessing the health of metallic structures using a microwave sensor depends on the reflection or/and transmission coefficient measurements as a function of the sweeping frequencies of the operating band. The aim of this work is reducing sweeping frequencies using machine learning feature selection techniques. By treating sweeping frequencies as features, the number of top important features can be identified, then only the most influential features (frequencies) are considered when building the microwave NDT equipment. The proposed method of reducing sweeping frequencies was validated experimentally using a waveguide sensor and a metallic plate with different cracks. Among the investigated feature selection techniques are information gain, gain ratio, relief, chi-squared. The effectiveness of the selected features were validated through performance evaluations of various classification models; namely, Nearest Neighbor, Neural Networks, Random Forest, and Support Vector Machine. Results showed good crack classification accuracy rates after employing feature selection algorithms. PMID:27104533

An audiovisual emotion recognition system

NASA Astrophysics Data System (ADS)

Han, Yi; Wang, Guoyin; Yang, Yong; He, Kun

2007-12-01

Human emotions could be expressed by many bio-symbols. Speech and facial expression are two of them. They are both regarded as emotional information which is playing an important role in human-computer interaction. Based on our previous studies on emotion recognition, an audiovisual emotion recognition system is developed and represented in this paper. The system is designed for real-time practice, and is guaranteed by some integrated modules. These modules include speech enhancement for eliminating noises, rapid face detection for locating face from background image, example based shape learning for facial feature alignment, and optical flow based tracking algorithm for facial feature tracking. It is known that irrelevant features and high dimensionality of the data can hurt the performance of classifier. Rough set-based feature selection is a good method for dimension reduction. So 13 speech features out of 37 ones and 10 facial features out of 33 ones are selected to represent emotional information, and 52 audiovisual features are selected due to the synchronization when speech and video fused together. The experiment results have demonstrated that this system performs well in real-time practice and has high recognition rate. Our results also show that the work in multimodules fused recognition will become the trend of emotion recognition in the future.

Respiratory trace feature analysis for the prediction of respiratory-gated PET quantification.

PubMed

Wang, Shouyi; Bowen, Stephen R; Chaovalitwongse, W Art; Sandison, George A; Grabowski, Thomas J; Kinahan, Paul E

2014-02-21

The benefits of respiratory gating in quantitative PET/CT vary tremendously between individual patients. Respiratory pattern is among many patient-specific characteristics that are thought to play an important role in gating-induced imaging improvements. However, the quantitative relationship between patient-specific characteristics of respiratory pattern and improvements in quantitative accuracy from respiratory-gated PET/CT has not been well established. If such a relationship could be estimated, then patient-specific respiratory patterns could be used to prospectively select appropriate motion compensation during image acquisition on a per-patient basis. This study was undertaken to develop a novel statistical model that predicts quantitative changes in PET/CT imaging due to respiratory gating. Free-breathing static FDG-PET images without gating and respiratory-gated FDG-PET images were collected from 22 lung and liver cancer patients on a PET/CT scanner. PET imaging quality was quantified with peak standardized uptake value (SUV(peak)) over lesions of interest. Relative differences in SUV(peak) between static and gated PET images were calculated to indicate quantitative imaging changes due to gating. A comprehensive multidimensional extraction of the morphological and statistical characteristics of respiratory patterns was conducted, resulting in 16 features that characterize representative patterns of a single respiratory trace. The six most informative features were subsequently extracted using a stepwise feature selection approach. The multiple-regression model was trained and tested based on a leave-one-subject-out cross-validation. The predicted quantitative improvements in PET imaging achieved an accuracy higher than 90% using a criterion with a dynamic error-tolerance range for SUV(peak) values. The results of this study suggest that our prediction framework could be applied to determine which patients would likely benefit from respiratory motion compensation when clinicians quantitatively assess PET/CT for therapy target definition and response assessment.

Respiratory trace feature analysis for the prediction of respiratory-gated PET quantification

NASA Astrophysics Data System (ADS)

Wang, Shouyi; Bowen, Stephen R.; Chaovalitwongse, W. Art; Sandison, George A.; Grabowski, Thomas J.; Kinahan, Paul E.

2014-02-01

The benefits of respiratory gating in quantitative PET/CT vary tremendously between individual patients. Respiratory pattern is among many patient-specific characteristics that are thought to play an important role in gating-induced imaging improvements. However, the quantitative relationship between patient-specific characteristics of respiratory pattern and improvements in quantitative accuracy from respiratory-gated PET/CT has not been well established. If such a relationship could be estimated, then patient-specific respiratory patterns could be used to prospectively select appropriate motion compensation during image acquisition on a per-patient basis. This study was undertaken to develop a novel statistical model that predicts quantitative changes in PET/CT imaging due to respiratory gating. Free-breathing static FDG-PET images without gating and respiratory-gated FDG-PET images were collected from 22 lung and liver cancer patients on a PET/CT scanner. PET imaging quality was quantified with peak standardized uptake value (SUVpeak) over lesions of interest. Relative differences in SUVpeak between static and gated PET images were calculated to indicate quantitative imaging changes due to gating. A comprehensive multidimensional extraction of the morphological and statistical characteristics of respiratory patterns was conducted, resulting in 16 features that characterize representative patterns of a single respiratory trace. The six most informative features were subsequently extracted using a stepwise feature selection approach. The multiple-regression model was trained and tested based on a leave-one-subject-out cross-validation. The predicted quantitative improvements in PET imaging achieved an accuracy higher than 90% using a criterion with a dynamic error-tolerance range for SUVpeak values. The results of this study suggest that our prediction framework could be applied to determine which patients would likely benefit from respiratory motion compensation when clinicians quantitatively assess PET/CT for therapy target definition and response assessment.

A Feature and Algorithm Selection Method for Improving the Prediction of Protein Structural Class.

PubMed

Ni, Qianwu; Chen, Lei

2017-01-01

Correct prediction of protein structural class is beneficial to investigation on protein functions, regulations and interactions. In recent years, several computational methods have been proposed in this regard. However, based on various features, it is still a great challenge to select proper classification algorithm and extract essential features to participate in classification. In this study, a feature and algorithm selection method was presented for improving the accuracy of protein structural class prediction. The amino acid compositions and physiochemical features were adopted to represent features and thirty-eight machine learning algorithms collected in Weka were employed. All features were first analyzed by a feature selection method, minimum redundancy maximum relevance (mRMR), producing a feature list. Then, several feature sets were constructed by adding features in the list one by one. For each feature set, thirtyeight algorithms were executed on a dataset, in which proteins were represented by features in the set. The predicted classes yielded by these algorithms and true class of each protein were collected to construct a dataset, which were analyzed by mRMR method, yielding an algorithm list. From the algorithm list, the algorithm was taken one by one to build an ensemble prediction model. Finally, we selected the ensemble prediction model with the best performance as the optimal ensemble prediction model. Experimental results indicate that the constructed model is much superior to models using single algorithm and other models that only adopt feature selection procedure or algorithm selection procedure. The feature selection procedure or algorithm selection procedure are really helpful for building an ensemble prediction model that can yield a better performance. Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.org.

Determining degree of optic nerve edema from color fundus photography

NASA Astrophysics Data System (ADS)

Agne, Jason; Wang, Jui-Kai; Kardon, Randy H.; Garvin, Mona K.

2015-03-01

Swelling of the optic nerve head (ONH) is subjectively assessed by clinicians using the Frisén scale. It is believed that a direct measurement of the ONH volume would serve as a better representation of the swelling. However, a direct measurement requires optic nerve imaging with spectral domain optical coherence tomography (SD-OCT) and 3D segmentation of the resulting images, which is not always available during clinical evaluation. Furthermore, telemedical imaging of the eye at remote locations is more feasible with non-mydriatic fundus cameras which are less costly than OCT imagers. Therefore, there is a critical need to develop a more quantitative analysis of optic nerve swelling on a continuous scale, similar to SD-OCT. Here, we select features from more commonly available 2D fundus images and use them to predict ONH volume. Twenty-six features were extracted from each of 48 color fundus images. The features include attributes of the blood vessels, optic nerve head, and peripapillary retina areas. These features were used in a regression analysis to predict ONH volume, as computed by a segmentation of the SD-OCT image. The results of the regression analysis yielded a mean square error of 2.43 mm3 and a correlation coefficient between computed and predicted volumes of R = 0:771, which suggests that ONH volume may be predicted from fundus features alone.

A feature-based approach to modeling protein-protein interaction hot spots.

PubMed

Cho, Kyu-il; Kim, Dongsup; Lee, Doheon

2009-05-01

Identifying features that effectively represent the energetic contribution of an individual interface residue to the interactions between proteins remains problematic. Here, we present several new features and show that they are more effective than conventional features. By combining the proposed features with conventional features, we develop a predictive model for interaction hot spots. Initially, 54 multifaceted features, composed of different levels of information including structure, sequence and molecular interaction information, are quantified. Then, to identify the best subset of features for predicting hot spots, feature selection is performed using a decision tree. Based on the selected features, a predictive model for hot spots is created using support vector machine (SVM) and tested on an independent test set. Our model shows better overall predictive accuracy than previous methods such as the alanine scanning methods Robetta and FOLDEF, and the knowledge-based method KFC. Subsequent analysis yields several findings about hot spots. As expected, hot spots have a larger relative surface area burial and are more hydrophobic than other residues. Unexpectedly, however, residue conservation displays a rather complicated tendency depending on the types of protein complexes, indicating that this feature is not good for identifying hot spots. Of the selected features, the weighted atomic packing density, relative surface area burial and weighted hydrophobicity are the top 3, with the weighted atomic packing density proving to be the most effective feature for predicting hot spots. Notably, we find that hot spots are closely related to pi-related interactions, especially pi . . . pi interactions.

Facial recognition using multisensor images based on localized kernel eigen spaces.

PubMed

Gundimada, Satyanadh; Asari, Vijayan K

2009-06-01

A feature selection technique along with an information fusion procedure for improving the recognition accuracy of a visual and thermal image-based facial recognition system is presented in this paper. A novel modular kernel eigenspaces approach is developed and implemented on the phase congruency feature maps extracted from the visual and thermal images individually. Smaller sub-regions from a predefined neighborhood within the phase congruency images of the training samples are merged to obtain a large set of features. These features are then projected into higher dimensional spaces using kernel methods. The proposed localized nonlinear feature selection procedure helps to overcome the bottlenecks of illumination variations, partial occlusions, expression variations and variations due to temperature changes that affect the visual and thermal face recognition techniques. AR and Equinox databases are used for experimentation and evaluation of the proposed technique. The proposed feature selection procedure has greatly improved the recognition accuracy for both the visual and thermal images when compared to conventional techniques. Also, a decision level fusion methodology is presented which along with the feature selection procedure has outperformed various other face recognition techniques in terms of recognition accuracy.

Feature Selection Methods for Robust Decoding of Finger Movements in a Non-human Primate

PubMed Central

Padmanaban, Subash; Baker, Justin; Greger, Bradley

2018-01-01

Objective: The performance of machine learning algorithms used for neural decoding of dexterous tasks may be impeded due to problems arising when dealing with high-dimensional data. The objective of feature selection algorithms is to choose a near-optimal subset of features from the original feature space to improve the performance of the decoding algorithm. The aim of our study was to compare the effects of four feature selection techniques, Wilcoxon signed-rank test, Relative Importance, Principal Component Analysis (PCA), and Mutual Information Maximization on SVM classification performance for a dexterous decoding task. Approach: A nonhuman primate (NHP) was trained to perform small coordinated movements—similar to typing. An array of microelectrodes was implanted in the hand area of the motor cortex of the NHP and used to record action potentials (AP) during finger movements. A Support Vector Machine (SVM) was used to classify which finger movement the NHP was making based upon AP firing rates. We used the SVM classification to examine the functional parameters of (i) robustness to simulated failure and (ii) longevity of classification. We also compared the effect of using isolated-neuron and multi-unit firing rates as the feature vector supplied to the SVM. Main results: The average decoding accuracy for multi-unit features and single-unit features using Mutual Information Maximization (MIM) across 47 sessions was 96.74 ± 3.5% and 97.65 ± 3.36% respectively. The reduction in decoding accuracy between using 100% of the features and 10% of features based on MIM was 45.56% (from 93.7 to 51.09%) and 4.75% (from 95.32 to 90.79%) for multi-unit and single-unit features respectively. MIM had best performance compared to other feature selection methods. Significance: These results suggest improved decoding performance can be achieved by using optimally selected features. The results based on clinically relevant performance metrics also suggest that the decoding algorithm can be made robust by using optimal features and feature selection algorithms. We believe that even a few percent increase in performance is important and improves the decoding accuracy of the machine learning algorithm potentially increasing the ease of use of a brain machine interface. PMID:29467602

Radiomics-based Prognosis Analysis for Non-Small Cell Lung Cancer

NASA Astrophysics Data System (ADS)

Zhang, Yucheng; Oikonomou, Anastasia; Wong, Alexander; Haider, Masoom A.; Khalvati, Farzad

2017-04-01

Radiomics characterizes tumor phenotypes by extracting large numbers of quantitative features from radiological images. Radiomic features have been shown to provide prognostic value in predicting clinical outcomes in several studies. However, several challenges including feature redundancy, unbalanced data, and small sample sizes have led to relatively low predictive accuracy. In this study, we explore different strategies for overcoming these challenges and improving predictive performance of radiomics-based prognosis for non-small cell lung cancer (NSCLC). CT images of 112 patients (mean age 75 years) with NSCLC who underwent stereotactic body radiotherapy were used to predict recurrence, death, and recurrence-free survival using a comprehensive radiomics analysis. Different feature selection and predictive modeling techniques were used to determine the optimal configuration of prognosis analysis. To address feature redundancy, comprehensive analysis indicated that Random Forest models and Principal Component Analysis were optimum predictive modeling and feature selection methods, respectively, for achieving high prognosis performance. To address unbalanced data, Synthetic Minority Over-sampling technique was found to significantly increase predictive accuracy. A full analysis of variance showed that data endpoints, feature selection techniques, and classifiers were significant factors in affecting predictive accuracy, suggesting that these factors must be investigated when building radiomics-based predictive models for cancer prognosis.

Normed kernel function-based fuzzy possibilistic C-means (NKFPCM) algorithm for high-dimensional breast cancer database classification with feature selection is based on Laplacian Score

NASA Astrophysics Data System (ADS)

Lestari, A. W.; Rustam, Z.

2017-07-01

In the last decade, breast cancer has become the focus of world attention as this disease is one of the primary leading cause of death for women. Therefore, it is necessary to have the correct precautions and treatment. In previous studies, Fuzzy Kennel K-Medoid algorithm has been used for multi-class data. This paper proposes an algorithm to classify the high dimensional data of breast cancer using Fuzzy Possibilistic C-means (FPCM) and a new method based on clustering analysis using Normed Kernel Function-Based Fuzzy Possibilistic C-Means (NKFPCM). The objective of this paper is to obtain the best accuracy in classification of breast cancer data. In order to improve the accuracy of the two methods, the features candidates are evaluated using feature selection, where Laplacian Score is used. The results show the comparison accuracy and running time of FPCM and NKFPCM with and without feature selection.

Identity Recognition Algorithm Using Improved Gabor Feature Selection of Gait Energy Image

NASA Astrophysics Data System (ADS)

Chao, LIANG; Ling-yao, JIA; Dong-cheng, SHI

2017-01-01

This paper describes an effective gait recognition approach based on Gabor features of gait energy image. In this paper, the kernel Fisher analysis combined with kernel matrix is proposed to select dominant features. The nearest neighbor classifier based on whitened cosine distance is used to discriminate different gait patterns. The approach proposed is tested on the CASIA and USF gait databases. The results show that our approach outperforms other state of gait recognition approaches in terms of recognition accuracy and robustness.

Combined texture feature analysis of segmentation and classification of benign and malignant tumour CT slices.

PubMed

Padma, A; Sukanesh, R

2013-01-01

A computer software system is designed for the segmentation and classification of benign from malignant tumour slices in brain computed tomography (CT) images. This paper presents a method to find and select both the dominant run length and co-occurrence texture features of region of interest (ROI) of the tumour region of each slice to be segmented by Fuzzy c means clustering (FCM) and evaluate the performance of support vector machine (SVM)-based classifiers in classifying benign and malignant tumour slices. Two hundred and six tumour confirmed CT slices are considered in this study. A total of 17 texture features are extracted by a feature extraction procedure, and six features are selected using Principal Component Analysis (PCA). This study constructed the SVM-based classifier with the selected features and by comparing the segmentation results with the experienced radiologist labelled ground truth (target). Quantitative analysis between ground truth and segmented tumour is presented in terms of segmentation accuracy, segmentation error and overlap similarity measures such as the Jaccard index. The classification performance of the SVM-based classifier with the same selected features is also evaluated using a 10-fold cross-validation method. The proposed system provides some newly found texture features have an important contribution in classifying benign and malignant tumour slices efficiently and accurately with less computational time. The experimental results showed that the proposed system is able to achieve the highest segmentation and classification accuracy effectiveness as measured by jaccard index and sensitivity and specificity.

Feature Selection Methods for Zero-Shot Learning of Neural Activity

PubMed Central

Caceres, Carlos A.; Roos, Matthew J.; Rupp, Kyle M.; Milsap, Griffin; Crone, Nathan E.; Wolmetz, Michael E.; Ratto, Christopher R.

2017-01-01

Dimensionality poses a serious challenge when making predictions from human neuroimaging data. Across imaging modalities, large pools of potential neural features (e.g., responses from particular voxels, electrodes, and temporal windows) have to be related to typically limited sets of stimuli and samples. In recent years, zero-shot prediction models have been introduced for mapping between neural signals and semantic attributes, which allows for classification of stimulus classes not explicitly included in the training set. While choices about feature selection can have a substantial impact when closed-set accuracy, open-set robustness, and runtime are competing design objectives, no systematic study of feature selection for these models has been reported. Instead, a relatively straightforward feature stability approach has been adopted and successfully applied across models and imaging modalities. To characterize the tradeoffs in feature selection for zero-shot learning, we compared correlation-based stability to several other feature selection techniques on comparable data sets from two distinct imaging modalities: functional Magnetic Resonance Imaging and Electrocorticography. While most of the feature selection methods resulted in similar zero-shot prediction accuracies and spatial/spectral patterns of selected features, there was one exception; A novel feature/attribute correlation approach was able to achieve those accuracies with far fewer features, suggesting the potential for simpler prediction models that yield high zero-shot classification accuracy. PMID:28690513

Human immunophenotyping via low-variance, low-bias, interpretive regression modeling of small, wide data sets: Application to aging and immune response to influenza vaccination.

PubMed

Holmes, Tyson H; He, Xiao-Song

2016-10-01

Small, wide data sets are commonplace in human immunophenotyping research. As defined here, a small, wide data set is constructed by sampling a small to modest quantity n,1

Human Immunophenotyping via Low-Variance, Low-Bias, Interpretive Regression Modeling of Small, Wide Data Sets: Application to Aging and Immune Response to Influenza Vaccination

PubMed Central

Holmes, Tyson H.; He, Xiao-Song

2016-01-01

Small, wide data sets are commonplace in human immunophenotyping research. As defined here, a small, wide data set is constructed by sampling a small to modest quantity n, 1 < n < 50, of human participants for the purpose of estimating many parameters p, such that n < p < 1,000. We offer a set of prescriptions that are designed to facilitate low-variance (i.e. stable), low-bias, interpretive regression modeling of small, wide data sets. These prescriptions are distinctive in their especially heavy emphasis on minimizing use of out-of-sample information for conducting statistical inference. That allows the working immunologist to proceed without being encumbered by imposed and often untestable statistical assumptions. Problems of unmeasured confounders, confidence-interval coverage, feature selection, and shrinkage/denoising are defined clearly and treated in detail. We propose an extension of an existing nonparametric technique for improved small-sample confidence-interval tail coverage from the univariate case (single immune feature) to the multivariate (many, possibly correlated immune features). An important role for derived features in the immunological interpretation of regression analyses is stressed. Areas of further research are discussed. Presented principles and methods are illustrated through application to a small, wide data set of adults spanning a wide range in ages and multiple immunophenotypes that were assayed before and after immunization with inactivated influenza vaccine (IIV). Our regression modeling prescriptions identify some potentially important topics for future immunological research. 1) Immunologists may wish to distinguish age-related differences in immune features from changes in immune features caused by aging. 2) A form of the bootstrap that employs linear extrapolation may prove to be an invaluable analytic tool because it allows the working immunologist to obtain accurate estimates of the stability of immune parameter estimates with a bare minimum of imposed assumptions. 3) Liberal inclusion of immune features in phenotyping panels can facilitate accurate separation of biological signal of interest from noise. In addition, through a combination of denoising and potentially improved confidence interval coverage, we identify some candidate immune correlates (frequency of cell subset and concentration of cytokine) with B cell response as measured by quantity of IIV-specific IgA antibody-secreting cells and quantity of IIV-specific IgG antibody-secreting cells. PMID:27196789

News video story segmentation method using fusion of audio-visual features

NASA Astrophysics Data System (ADS)

Wen, Jun; Wu, Ling-da; Zeng, Pu; Luan, Xi-dao; Xie, Yu-xiang

2007-11-01

News story segmentation is an important aspect for news video analysis. This paper presents a method for news video story segmentation. Different form prior works, which base on visual features transform, the proposed technique uses audio features as baseline and fuses visual features with it to refine the results. At first, it selects silence clips as audio features candidate points, and selects shot boundaries and anchor shots as two kinds of visual features candidate points. Then this paper selects audio feature candidates as cues and develops different fusion method, which effectively using diverse type visual candidates to refine audio candidates, to get story boundaries. Experiment results show that this method has high efficiency and adaptability to different kinds of news video.

An opinion formation based binary optimization approach for feature selection

NASA Astrophysics Data System (ADS)

Hamedmoghadam, Homayoun; Jalili, Mahdi; Yu, Xinghuo

2018-02-01

This paper proposed a novel optimization method based on opinion formation in complex network systems. The proposed optimization technique mimics human-human interaction mechanism based on a mathematical model derived from social sciences. Our method encodes a subset of selected features to the opinion of an artificial agent and simulates the opinion formation process among a population of agents to solve the feature selection problem. The agents interact using an underlying interaction network structure and get into consensus in their opinions, while finding better solutions to the problem. A number of mechanisms are employed to avoid getting trapped in local minima. We compare the performance of the proposed method with a number of classical population-based optimization methods and a state-of-the-art opinion formation based method. Our experiments on a number of high dimensional datasets reveal outperformance of the proposed algorithm over others.

A semisupervised support vector regression method to estimate biophysical parameters from remotely sensed images

NASA Astrophysics Data System (ADS)

Castelletti, Davide; Demir, Begüm; Bruzzone, Lorenzo

2014-10-01

This paper presents a novel semisupervised learning (SSL) technique defined in the context of ɛ-insensitive support vector regression (SVR) to estimate biophysical parameters from remotely sensed images. The proposed SSL method aims to mitigate the problems of small-sized biased training sets without collecting any additional samples with reference measures. This is achieved on the basis of two consecutive steps. The first step is devoted to inject additional priors information in the learning phase of the SVR in order to adapt the importance of each training sample according to distribution of the unlabeled samples. To this end, a weight is initially associated to each training sample based on a novel strategy that defines higher weights for the samples located in the high density regions of the feature space while giving reduced weights to those that fall into the low density regions of the feature space. Then, in order to exploit different weights for training samples in the learning phase of the SVR, we introduce a weighted SVR (WSVR) algorithm. The second step is devoted to jointly exploit labeled and informative unlabeled samples for further improving the definition of the WSVR learning function. To this end, the most informative unlabeled samples that have an expected accurate target values are initially selected according to a novel strategy that relies on the distribution of the unlabeled samples in the feature space and on the WSVR function estimated at the first step. Then, we introduce a restructured WSVR algorithm that jointly uses labeled and unlabeled samples in the learning phase of the WSVR algorithm and tunes their importance by different values of regularization parameters. Experimental results obtained for the estimation of single-tree stem volume show the effectiveness of the proposed SSL method.

Text-Based Conferencing: Features vs. Functionality

ERIC Educational Resources Information Center

Anderson, Lynn; McCarthy, Cathy

2005-01-01

This report examines three text-based conferencing products: "WowBB", "Invision Power Board", and "vBulletin". Their selection was prompted by a feature-by-feature comparison of the same products on the "WowBB" website. The comparison chart painted a misleading impression of "WowBB's" features in relation to the other two products; so the…

HOS network-based classification of power quality events via regression algorithms

NASA Astrophysics Data System (ADS)

Palomares Salas, José Carlos; González de la Rosa, Juan José; Sierra Fernández, José María; Pérez, Agustín Agüera

2015-12-01

This work compares seven regression algorithms implemented in artificial neural networks (ANNs) supported by 14 power-quality features, which are based in higher-order statistics. Combining time and frequency domain estimators to deal with non-stationary measurement sequences, the final goal of the system is the implementation in the future smart grid to guarantee compatibility between all equipment connected. The principal results are based in spectral kurtosis measurements, which easily adapt to the impulsive nature of the power quality events. These results verify that the proposed technique is capable of offering interesting results for power quality (PQ) disturbance classification. The best results are obtained using radial basis networks, generalized regression, and multilayer perceptron, mainly due to the non-linear nature of data.

Unsupervised Feature Selection Based on the Morisita Index for Hyperspectral Images

NASA Astrophysics Data System (ADS)

Golay, Jean; Kanevski, Mikhail

2017-04-01

Hyperspectral sensors are capable of acquiring images with hundreds of narrow and contiguous spectral bands. Compared with traditional multispectral imagery, the use of hyperspectral images allows better performance in discriminating between land-cover classes, but it also results in large redundancy and high computational data processing. To alleviate such issues, unsupervised feature selection techniques for redundancy minimization can be implemented. Their goal is to select the smallest subset of features (or bands) in such a way that all the information content of a data set is preserved as much as possible. The present research deals with the application to hyperspectral images of a recently introduced technique of unsupervised feature selection: the Morisita-Based filter for Redundancy Minimization (MBRM). MBRM is based on the (multipoint) Morisita index of clustering and on the Morisita estimator of Intrinsic Dimension (ID). The fundamental idea of the technique is to retain only the bands which contribute to increasing the ID of an image. In this way, redundant bands are disregarded, since they have no impact on the ID. Besides, MBRM has several advantages over benchmark techniques: in addition to its ability to deal with large data sets, it can capture highly-nonlinear dependences and its implementation is straightforward in any programming environment. Experimental results on freely available hyperspectral images show the good effectiveness of MBRM in remote sensing data processing. Comparisons with benchmark techniques are carried out and random forests are used to assess the performance of MBRM in reducing the data dimensionality without loss of relevant information. References [1] C. Traina Jr., A.J.M. Traina, L. Wu, C. Faloutsos, Fast feature selection using fractal dimension, in: Proceedings of the XV Brazilian Symposium on Databases, SBBD, pp. 158-171, 2000. [2] J. Golay, M. Kanevski, A new estimator of intrinsic dimension based on the multipoint Morisita index, Pattern Recognition 48(12), pp. 4070-4081, 2015. [3] J. Golay, M. Kanevski, Unsupervised feature selection based on the Morisita estimator of intrinsic dimension, arXiv:1608.05581, 2016.

Development and selection of Asian-specific humeral implants based on statistical atlas: toward planning minimally invasive surgery.

PubMed

Wu, K; Daruwalla, Z J; Wong, K L; Murphy, D; Ren, H

2015-08-01

The commercial humeral implants based on the Western population are currently not entirely compatible with Asian patients, due to differences in bone size, shape and structure. Surgeons may have to compromise or use different implants that are less conforming, which may cause complications of as well as inconvenience to the implant position. The construction of Asian humerus atlases of different clusters has therefore been proposed to eradicate this problem and to facilitate planning minimally invasive surgical procedures [6,31]. According to the features of the atlases, new implants could be designed specifically for different patients. Furthermore, an automatic implant selection algorithm has been proposed as well in order to reduce the complications caused by implant and bone mismatch. Prior to the design of the implant, data clustering and extraction of the relevant features were carried out on the datasets of each gender. The fuzzy C-means clustering method is explored in this paper. Besides, two new schemes of implant selection procedures, namely the Procrustes analysis-based scheme and the group average distance-based scheme, were proposed to better search for the matching implants for new coming patients from the database. Both these two algorithms have not been used in this area, while they turn out to have excellent performance in implant selection. Additionally, algorithms to calculate the matching scores between various implants and the patient data are proposed in this paper to assist the implant selection procedure. The results obtained have indicated the feasibility of the proposed development and selection scheme. The 16 sets of male data were divided into two clusters with 8 and 8 subjects, respectively, and the 11 female datasets were also divided into two clusters with 5 and 6 subjects, respectively. Based on the features of each cluster, the implants designed by the proposed algorithm fit very well on their reference humeri and the proposed implant selection procedure allows for a scenario of treating a patient with merely a preoperative anatomical model in order to correctly select the implant that has the best fit. Based on the leave-one-out validation, it can be concluded that both the PA-based method and GAD-based method are able to achieve excellent performance when dealing with the problem of implant selection. The accuracy and average execution time for the PA-based method were 100 % and 0.132 s, respectively, while those of the GAD- based method were 100 % and 0.058 s. Therefore, the GAD-based method outperformed the PA-based method in terms of execution speed. The primary contributions of this paper include the proposal of methods for development of Asian-, gender- and cluster-specific implants based on shape features and selection of the best fit implants for future patients according to their features. To the best of our knowledge, this is the first work that proposes implant design and selection for Asian patients automatically based on features extracted from cluster-specific statistical atlases.

A New Approach to Develop Computer-aided Diagnosis Scheme of Breast Mass Classification Using Deep Learning Technology

PubMed Central

Qiu, Yuchen; Yan, Shiju; Gundreddy, Rohith Reddy; Wang, Yunzhi; Cheng, Samuel; Liu, Hong; Zheng, Bin

2017-01-01

PURPOSE To develop and test a deep learning based computer-aided diagnosis (CAD) scheme of mammograms for classifying between malignant and benign masses. METHODS An image dataset involving 560 regions of interest (ROIs) extracted from digital mammograms was used. After down-sampling each ROI from 512×512 to 64×64 pixel size, we applied an 8 layer deep learning network that involves 3 pairs of convolution-max-pooling layers for automatic feature extraction and a multiple layer perceptron (MLP) classifier for feature categorization to process ROIs. The 3 pairs of convolution layers contain 20, 10, and 5 feature maps, respectively. Each convolution layer is connected with a max-pooling layer to improve the feature robustness. The output of the sixth layer is fully connected with a MLP classifier, which is composed of one hidden layer and one logistic regression layer. The network then generates a classification score to predict the likelihood of ROI depicting a malignant mass. A four-fold cross validation method was applied to train and test this deep learning network. RESULTS The results revealed that this CAD scheme yields an area under the receiver operation characteristic curve (AUC) of 0.696±0.044, 0.802±0.037, 0.836±0.036, and 0.822±0.035 for fold 1 to 4 testing datasets, respectively. The overall AUC of the entire dataset is 0.790±0.019. CONCLUSIONS This study demonstrates the feasibility of applying a deep learning based CAD scheme to classify between malignant and benign breast masses without a lesion segmentation, image feature computation and selection process. PMID:28436410

A new approach to develop computer-aided diagnosis scheme of breast mass classification using deep learning technology.

PubMed

Qiu, Yuchen; Yan, Shiju; Gundreddy, Rohith Reddy; Wang, Yunzhi; Cheng, Samuel; Liu, Hong; Zheng, Bin

2017-01-01

To develop and test a deep learning based computer-aided diagnosis (CAD) scheme of mammograms for classifying between malignant and benign masses. An image dataset involving 560 regions of interest (ROIs) extracted from digital mammograms was used. After down-sampling each ROI from 512×512 to 64×64 pixel size, we applied an 8 layer deep learning network that involves 3 pairs of convolution-max-pooling layers for automatic feature extraction and a multiple layer perceptron (MLP) classifier for feature categorization to process ROIs. The 3 pairs of convolution layers contain 20, 10, and 5 feature maps, respectively. Each convolution layer is connected with a max-pooling layer to improve the feature robustness. The output of the sixth layer is fully connected with a MLP classifier, which is composed of one hidden layer and one logistic regression layer. The network then generates a classification score to predict the likelihood of ROI depicting a malignant mass. A four-fold cross validation method was applied to train and test this deep learning network. The results revealed that this CAD scheme yields an area under the receiver operation characteristic curve (AUC) of 0.696±0.044, 0.802±0.037, 0.836±0.036, and 0.822±0.035 for fold 1 to 4 testing datasets, respectively. The overall AUC of the entire dataset is 0.790±0.019. This study demonstrates the feasibility of applying a deep learning based CAD scheme to classify between malignant and benign breast masses without a lesion segmentation, image feature computation and selection process.

Fuzzy feature selection based on interval type-2 fuzzy sets

NASA Astrophysics Data System (ADS)

Cherif, Sahar; Baklouti, Nesrine; Alimi, Adel; Snasel, Vaclav

2017-03-01

When dealing with real world data; noise, complexity, dimensionality, uncertainty and irrelevance can lead to low performance and insignificant judgment. Fuzzy logic is a powerful tool for controlling conflicting attributes which can have similar effects and close meanings. In this paper, an interval type-2 fuzzy feature selection is presented as a new approach for removing irrelevant features and reducing complexity. We demonstrate how can Feature Selection be joined with Interval Type-2 Fuzzy Logic for keeping significant features and hence reducing time complexity. The proposed method is compared with some other approaches. The results show that the number of attributes is proportionally small.

PrAS: Prediction of amidation sites using multiple feature extraction.

PubMed

Wang, Tong; Zheng, Wei; Wuyun, Qiqige; Wu, Zhenfeng; Ruan, Jishou; Hu, Gang; Gao, Jianzhao

2017-02-01

Amidation plays an important role in a variety of pathological processes and serious diseases like neural dysfunction and hypertension. However, identification of protein amidation sites through traditional experimental methods is time consuming and expensive. In this paper, we proposed a novel predictor for Prediction of Amidation Sites (PrAS), which is the first software package for academic users. The method incorporated four representative feature types, which are position-based features, physicochemical and biochemical properties features, predicted structure-based features and evolutionary information features. A novel feature selection method, positive contribution feature selection was proposed to optimize features. PrAS achieved AUC of 0.96, accuracy of 92.1%, sensitivity of 81.2%, specificity of 94.9% and MCC of 0.76 on the independent test set. PrAS is freely available at https://sourceforge.net/p/praspkg. Copyright © 2016 Elsevier Ltd. All rights reserved.

EXpectation Propagation LOgistic REgRession (EXPLORER): distributed privacy-preserving online model learning.

PubMed

Wang, Shuang; Jiang, Xiaoqian; Wu, Yuan; Cui, Lijuan; Cheng, Samuel; Ohno-Machado, Lucila

2013-06-01

We developed an EXpectation Propagation LOgistic REgRession (EXPLORER) model for distributed privacy-preserving online learning. The proposed framework provides a high level guarantee for protecting sensitive information, since the information exchanged between the server and the client is the encrypted posterior distribution of coefficients. Through experimental results, EXPLORER shows the same performance (e.g., discrimination, calibration, feature selection, etc.) as the traditional frequentist logistic regression model, but provides more flexibility in model updating. That is, EXPLORER can be updated one point at a time rather than having to retrain the entire data set when new observations are recorded. The proposed EXPLORER supports asynchronized communication, which relieves the participants from coordinating with one another, and prevents service breakdown from the absence of participants or interrupted communications. Copyright © 2013 Elsevier Inc. All rights reserved.

The role of lightness, hue and saturation in feature-based visual attention.

PubMed

Stuart, Geoffrey W; Barsdell, Wendy N; Day, Ross H

2014-03-01

Visual attention is used to select part of the visual array for higher-level processing. Visual selection can be based on spatial location, but it has also been demonstrated that multiple locations can be selected simultaneously on the basis of a visual feature such as color. One task that has been used to demonstrate feature-based attention is the judgement of the symmetry of simple four-color displays. In a typical task, when symmetry is violated, four squares on either side of the display do not match. When four colors are involved, symmetry judgements are made more quickly than when only two of the four colors are involved. This indicates that symmetry judgements are made one color at a time. Previous studies have confounded lightness, hue, and saturation when defining the colors used in such displays. In three experiments, symmetry was defined by lightness alone, lightness plus hue, or by hue or saturation alone, with lightness levels randomised. The difference between judgements of two- and four-color asymmetry was maintained, showing that hue and saturation can provide the sole basis for feature-based attentional selection. Crown Copyright © 2014. Published by Elsevier Ltd. All rights reserved.

A Genetic-Based Feature Selection Approach in the Identification of Left/Right Hand Motor Imagery for a Brain-Computer Interface

PubMed Central

Yaacoub, Charles; Mhanna, Georges; Rihana, Sandy

2017-01-01

Electroencephalography is a non-invasive measure of the brain electrical activity generated by millions of neurons. Feature extraction in electroencephalography analysis is a core issue that may lead to accurate brain mental state classification. This paper presents a new feature selection method that improves left/right hand movement identification of a motor imagery brain-computer interface, based on genetic algorithms and artificial neural networks used as classifiers. Raw electroencephalography signals are first preprocessed using appropriate filtering. Feature extraction is carried out afterwards, based on spectral and temporal signal components, and thus a feature vector is constructed. As various features might be inaccurate and mislead the classifier, thus degrading the overall system performance, the proposed approach identifies a subset of features from a large feature space, such that the classifier error rate is reduced. Experimental results show that the proposed method is able to reduce the number of features to as low as 0.5% (i.e., the number of ignored features can reach 99.5%) while improving the accuracy, sensitivity, specificity, and precision of the classifier. PMID:28124985