Regional regression of flood characteristics employing historical information

USGS Publications Warehouse

Tasker, Gary D.; Stedinger, J.R.

1987-01-01

Streamflow gauging networks provide hydrologic information for use in estimating the parameters of regional regression models. The regional regression models can be used to estimate flood statistics, such as the 100 yr peak, at ungauged sites as functions of drainage basin characteristics. A recent innovation in regional regression is the use of a generalized least squares (GLS) estimator that accounts for unequal station record lengths and sample cross correlation among the flows. However, this technique does not account for historical flood information. A method is proposed here to adjust this generalized least squares estimator to account for possible information about historical floods available at some stations in a region. The historical information is assumed to be in the form of observations of all peaks above a threshold during a long period outside the systematic record period. A Monte Carlo simulation experiment was performed to compare the GLS estimator adjusted for historical floods with the unadjusted GLS estimator and the ordinary least squares estimator. Results indicate that using the GLS estimator adjusted for historical information significantly improves the regression model. ?? 1987.

Exact Analysis of Squared Cross-Validity Coefficient in Predictive Regression Models

ERIC Educational Resources Information Center

Shieh, Gwowen

2009-01-01

In regression analysis, the notion of population validity is of theoretical interest for describing the usefulness of the underlying regression model, whereas the presumably more important concept of population cross-validity represents the predictive effectiveness for the regression equation in future research. It appears that the inference…

Methods for estimating the magnitude and frequency of peak streamflows at ungaged sites in and near the Oklahoma Panhandle

USGS Publications Warehouse

Smith, S. Jerrod; Lewis, Jason M.; Graves, Grant M.

2015-09-28

Generalized-least-squares multiple-linear regression analysis was used to formulate regression relations between peak-streamflow frequency statistics and basin characteristics. Contributing drainage area was the only basin characteristic determined to be statistically significant for all percentage of annual exceedance probabilities and was the only basin characteristic used in regional regression equations for estimating peak-streamflow frequency statistics on unregulated streams in and near the Oklahoma Panhandle. The regression model pseudo-coefficient of determination, converted to percent, for the Oklahoma Panhandle regional regression equations ranged from about 38 to 63 percent. The standard errors of prediction and the standard model errors for the Oklahoma Panhandle regional regression equations ranged from about 84 to 148 percent and from about 76 to 138 percent, respectively. These errors were comparable to those reported for regional peak-streamflow frequency regression equations for the High Plains areas of Texas and Colorado. The root mean square errors for the Oklahoma Panhandle regional regression equations (ranging from 3,170 to 92,000 cubic feet per second) were less than the root mean square errors for the Oklahoma statewide regression equations (ranging from 18,900 to 412,000 cubic feet per second); therefore, the Oklahoma Panhandle regional regression equations produce more accurate peak-streamflow statistic estimates for the irrigated period of record in the Oklahoma Panhandle than do the Oklahoma statewide regression equations. The regression equations developed in this report are applicable to streams that are not substantially affected by regulation, impoundment, or surface-water withdrawals. These regression equations are intended for use for stream sites with contributing drainage areas less than or equal to about 2,060 square miles, the maximum value for the independent variable used in the regression analysis.

Estimating current and future streamflow characteristics at ungaged sites, central and eastern Montana, with application to evaluating effects of climate change on fish populations

USGS Publications Warehouse

Sando, Roy; Chase, Katherine J.

2017-03-23

A common statistical procedure for estimating streamflow statistics at ungaged locations is to develop a relational model between streamflow and drainage basin characteristics at gaged locations using least squares regression analysis; however, least squares regression methods are parametric and make constraining assumptions about the data distribution. The random forest regression method provides an alternative nonparametric method for estimating streamflow characteristics at ungaged sites and requires that the data meet fewer statistical conditions than least squares regression methods.Random forest regression analysis was used to develop predictive models for 89 streamflow characteristics using Precipitation-Runoff Modeling System simulated streamflow data and drainage basin characteristics at 179 sites in central and eastern Montana. The predictive models were developed from streamflow data simulated for current (baseline, water years 1982–99) conditions and three future periods (water years 2021–38, 2046–63, and 2071–88) under three different climate-change scenarios. These predictive models were then used to predict streamflow characteristics for baseline conditions and three future periods at 1,707 fish sampling sites in central and eastern Montana. The average root mean square error for all predictive models was about 50 percent. When streamflow predictions at 23 fish sampling sites were compared to nearby locations with simulated data, the mean relative percent difference was about 43 percent. When predictions were compared to streamflow data recorded at 21 U.S. Geological Survey streamflow-gaging stations outside of the calibration basins, the average mean absolute percent error was about 73 percent.

Geodesic regression on orientation distribution functions with its application to an aging study.

PubMed

Du, Jia; Goh, Alvina; Kushnarev, Sergey; Qiu, Anqi

2014-02-15

In this paper, we treat orientation distribution functions (ODFs) derived from high angular resolution diffusion imaging (HARDI) as elements of a Riemannian manifold and present a method for geodesic regression on this manifold. In order to find the optimal regression model, we pose this as a least-squares problem involving the sum-of-squared geodesic distances between observed ODFs and their model fitted data. We derive the appropriate gradient terms and employ gradient descent to find the minimizer of this least-squares optimization problem. In addition, we show how to perform statistical testing for determining the significance of the relationship between the manifold-valued regressors and the real-valued regressands. Experiments on both synthetic and real human data are presented. In particular, we examine aging effects on HARDI via geodesic regression of ODFs in normal adults aged 22 years old and above. © 2013 Elsevier Inc. All rights reserved.

A consensus least squares support vector regression (LS-SVR) for analysis of near-infrared spectra of plant samples.

PubMed

Li, Yankun; Shao, Xueguang; Cai, Wensheng

2007-04-15

Consensus modeling of combining the results of multiple independent models to produce a single prediction avoids the instability of single model. Based on the principle of consensus modeling, a consensus least squares support vector regression (LS-SVR) method for calibrating the near-infrared (NIR) spectra was proposed. In the proposed approach, NIR spectra of plant samples were firstly preprocessed using discrete wavelet transform (DWT) for filtering the spectral background and noise, then, consensus LS-SVR technique was used for building the calibration model. With an optimization of the parameters involved in the modeling, a satisfied model was achieved for predicting the content of reducing sugar in plant samples. The predicted results show that consensus LS-SVR model is more robust and reliable than the conventional partial least squares (PLS) and LS-SVR methods.

Geodesic least squares regression for scaling studies in magnetic confinement fusion

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

Verdoolaege, Geert

In regression analyses for deriving scaling laws that occur in various scientific disciplines, usually standard regression methods have been applied, of which ordinary least squares (OLS) is the most popular. However, concerns have been raised with respect to several assumptions underlying OLS in its application to scaling laws. We here discuss a new regression method that is robust in the presence of significant uncertainty on both the data and the regression model. The method, which we call geodesic least squares regression (GLS), is based on minimization of the Rao geodesic distance on a probabilistic manifold. We demonstrate the superiority ofmore » the method using synthetic data and we present an application to the scaling law for the power threshold for the transition to the high confinement regime in magnetic confinement fusion devices.« less

Superquantile Regression: Theory, Algorithms, and Applications

DTIC Science & Technology

2014-12-01

Example C: Stack loss data scatterplot matrix. 91 Regression α c0 caf cwt cac R̄ 2 α R̄ 2 α,Adj Least Squares NA -39.9197 0.7156 1.2953 -0.1521 0.9136...This is due to a small 92 Model Regression α c0 cwt cwt2 R̄ 2 α R̄ 2 α,Adj f2 Least Squares NA -41.9109 2.8174 — 0.7665 0.7542 Quantile 0.25 -32.0000

Rapid Detection of Volatile Oil in Mentha haplocalyx by Near-Infrared Spectroscopy and Chemometrics.

PubMed

Yan, Hui; Guo, Cheng; Shao, Yang; Ouyang, Zhen

2017-01-01

Near-infrared spectroscopy combined with partial least squares regression (PLSR) and support vector machine (SVM) was applied for the rapid determination of chemical component of volatile oil content in Mentha haplocalyx . The effects of data pre-processing methods on the accuracy of the PLSR calibration models were investigated. The performance of the final model was evaluated according to the correlation coefficient ( R ) and root mean square error of prediction (RMSEP). For PLSR model, the best preprocessing method combination was first-order derivative, standard normal variate transformation (SNV), and mean centering, which had of 0.8805, of 0.8719, RMSEC of 0.091, and RMSEP of 0.097, respectively. The wave number variables linking to volatile oil are from 5500 to 4000 cm-1 by analyzing the loading weights and variable importance in projection (VIP) scores. For SVM model, six LVs (less than seven LVs in PLSR model) were adopted in model, and the result was better than PLSR model. The and were 0.9232 and 0.9202, respectively, with RMSEC and RMSEP of 0.084 and 0.082, respectively, which indicated that the predicted values were accurate and reliable. This work demonstrated that near infrared reflectance spectroscopy with chemometrics could be used to rapidly detect the main content volatile oil in M. haplocalyx . The quality of medicine directly links to clinical efficacy, thus, it is important to control the quality of Mentha haplocalyx . Near-infrared spectroscopy combined with partial least squares regression (PLSR) and support vector machine (SVM) was applied for the rapid determination of chemical component of volatile oil content in Mentha haplocalyx . For SVM model, 6 LVs (less than 7 LVs in PLSR model) were adopted in model, and the result was better than PLSR model. It demonstrated that near infrared reflectance spectroscopy with chemometrics could be used to rapidly detect the main content volatile oil in Mentha haplocalyx . Abbreviations used: 1 st der: First-order derivative; 2 nd der: Second-order derivative; LOO: Leave-one-out; LVs: Latent variables; MC: Mean centering, NIR: Near-infrared; NIRS: Near infrared spectroscopy; PCR: Principal component regression, PLSR: Partial least squares regression; RBF: Radial basis function; RMSEC: Root mean square error of cross validation, RMSEC: Root mean square error of calibration; RMSEP: Root mean square error of prediction; SNV: Standard normal variate transformation; SVM: Support vector machine; VIP: Variable Importance in projection.

Weighted linear regression using D2H and D2 as the independent variables

Treesearch

Hans T. Schreuder; Michael S. Williams

1998-01-01

Several error structures for weighted regression equations used for predicting volume were examined for 2 large data sets of felled and standing loblolly pine trees (Pinus taeda L.). The generally accepted model with variance of error proportional to the value of the covariate squared ( D2H = diameter squared times height or D...

Using partial least squares regression as a predictive tool in describing equine third metacarpal bone shape.

PubMed

Liley, Helen; Zhang, Ju; Firth, Elwyn; Fernandez, Justin; Besier, Thor

2017-11-01

Population variance in bone shape is an important consideration when applying the results of subject-specific computational models to a population. In this letter, we demonstrate the ability of partial least squares regression to provide an improved shape prediction of the equine third metacarpal epiphysis, using two easily obtained measurements.

Inverse models: A necessary next step in ground-water modeling

USGS Publications Warehouse

Poeter, E.P.; Hill, M.C.

1997-01-01

Inverse models using, for example, nonlinear least-squares regression, provide capabilities that help modelers take full advantage of the insight available from ground-water models. However, lack of information about the requirements and benefits of inverse models is an obstacle to their widespread use. This paper presents a simple ground-water flow problem to illustrate the requirements and benefits of the nonlinear least-squares repression method of inverse modeling and discusses how these attributes apply to field problems. The benefits of inverse modeling include: (1) expedited determination of best fit parameter values; (2) quantification of the (a) quality of calibration, (b) data shortcomings and needs, and (c) confidence limits on parameter estimates and predictions; and (3) identification of issues that are easily overlooked during nonautomated calibration.Inverse models using, for example, nonlinear least-squares regression, provide capabilities that help modelers take full advantage of the insight available from ground-water models. However, lack of information about the requirements and benefits of inverse models is an obstacle to their widespread use. This paper presents a simple ground-water flow problem to illustrate the requirements and benefits of the nonlinear least-squares regression method of inverse modeling and discusses how these attributes apply to field problems. The benefits of inverse modeling include: (1) expedited determination of best fit parameter values; (2) quantification of the (a) quality of calibration, (b) data shortcomings and needs, and (c) confidence limits on parameter estimates and predictions; and (3) identification of issues that are easily overlooked during nonautomated calibration.

Assessment of parametric uncertainty for groundwater reactive transport modeling,

USGS Publications Warehouse

Shi, Xiaoqing; Ye, Ming; Curtis, Gary P.; Miller, Geoffery L.; Meyer, Philip D.; Kohler, Matthias; Yabusaki, Steve; Wu, Jichun

2014-01-01

The validity of using Gaussian assumptions for model residuals in uncertainty quantification of a groundwater reactive transport model was evaluated in this study. Least squares regression methods explicitly assume Gaussian residuals, and the assumption leads to Gaussian likelihood functions, model parameters, and model predictions. While the Bayesian methods do not explicitly require the Gaussian assumption, Gaussian residuals are widely used. This paper shows that the residuals of the reactive transport model are non-Gaussian, heteroscedastic, and correlated in time; characterizing them requires using a generalized likelihood function such as the formal generalized likelihood function developed by Schoups and Vrugt (2010). For the surface complexation model considered in this study for simulating uranium reactive transport in groundwater, parametric uncertainty is quantified using the least squares regression methods and Bayesian methods with both Gaussian and formal generalized likelihood functions. While the least squares methods and Bayesian methods with Gaussian likelihood function produce similar Gaussian parameter distributions, the parameter distributions of Bayesian uncertainty quantification using the formal generalized likelihood function are non-Gaussian. In addition, predictive performance of formal generalized likelihood function is superior to that of least squares regression and Bayesian methods with Gaussian likelihood function. The Bayesian uncertainty quantification is conducted using the differential evolution adaptive metropolis (DREAM(zs)) algorithm; as a Markov chain Monte Carlo (MCMC) method, it is a robust tool for quantifying uncertainty in groundwater reactive transport models. For the surface complexation model, the regression-based local sensitivity analysis and Morris- and DREAM(ZS)-based global sensitivity analysis yield almost identical ranking of parameter importance. The uncertainty analysis may help select appropriate likelihood functions, improve model calibration, and reduce predictive uncertainty in other groundwater reactive transport and environmental modeling.

Addressing the identification problem in age-period-cohort analysis: a tutorial on the use of partial least squares and principal components analysis.

PubMed

Tu, Yu-Kang; Krämer, Nicole; Lee, Wen-Chung

2012-07-01

In the analysis of trends in health outcomes, an ongoing issue is how to separate and estimate the effects of age, period, and cohort. As these 3 variables are perfectly collinear by definition, regression coefficients in a general linear model are not unique. In this tutorial, we review why identification is a problem, and how this problem may be tackled using partial least squares and principal components regression analyses. Both methods produce regression coefficients that fulfill the same collinearity constraint as the variables age, period, and cohort. We show that, because the constraint imposed by partial least squares and principal components regression is inherent in the mathematical relation among the 3 variables, this leads to more interpretable results. We use one dataset from a Taiwanese health-screening program to illustrate how to use partial least squares regression to analyze the trends in body heights with 3 continuous variables for age, period, and cohort. We then use another dataset of hepatocellular carcinoma mortality rates for Taiwanese men to illustrate how to use partial least squares regression to analyze tables with aggregated data. We use the second dataset to show the relation between the intrinsic estimator, a recently proposed method for the age-period-cohort analysis, and partial least squares regression. We also show that the inclusion of all indicator variables provides a more consistent approach. R code for our analyses is provided in the eAppendix.

Fourier transform infrared reflectance spectra of latent fingerprints: a biometric gauge for the age of an individual.

PubMed

Hemmila, April; McGill, Jim; Ritter, David

2008-03-01

To determine if changes in fingerprint infrared spectra linear with age can be found, partial least squares (PLS1) regression of 155 fingerprint infrared spectra against the person's age was constructed. The regression produced a linear model of age as a function of spectrum with a root mean square error of calibration of less than 4 years, showing an inflection at about 25 years of age. The spectral ranges emphasized by the regression do not correspond to the highest concentration constituents of the fingerprints. Separate linear regression models for old and young people can be constructed with even more statistical rigor. The success of the regression demonstrates that a combination of constituents can be found that changes linearly with age, with a significant shift around puberty.

Incorporation of prior information on parameters into nonlinear regression groundwater flow models: 2. Applications

USGS Publications Warehouse

Cooley, Richard L.

1983-01-01

This paper investigates factors influencing the degree of improvement in estimates of parameters of a nonlinear regression groundwater flow model by incorporating prior information of unknown reliability. Consideration of expected behavior of the regression solutions and results of a hypothetical modeling problem lead to several general conclusions. First, if the parameters are properly scaled, linearized expressions for the mean square error (MSE) in parameter estimates of a nonlinear model will often behave very nearly as if the model were linear. Second, by using prior information, the MSE in properly scaled parameters can be reduced greatly over the MSE of ordinary least squares estimates of parameters. Third, plots of estimated MSE and the estimated standard deviation of MSE versus an auxiliary parameter (the ridge parameter) specifying the degree of influence of the prior information on regression results can help determine the potential for improvement of parameter estimates. Fourth, proposed criteria can be used to make appropriate choices for the ridge parameter and another parameter expressing degree of overall bias in the prior information. Results of a case study of Truckee Meadows, Reno-Sparks area, Washoe County, Nevada, conform closely to the results of the hypothetical problem. In the Truckee Meadows case, incorporation of prior information did not greatly change the parameter estimates from those obtained by ordinary least squares. However, the analysis showed that both sets of estimates are more reliable than suggested by the standard errors from ordinary least squares.

Applicability of Monte Carlo cross validation technique for model development and validation using generalised least squares regression

NASA Astrophysics Data System (ADS)

Haddad, Khaled; Rahman, Ataur; A Zaman, Mohammad; Shrestha, Surendra

2013-03-01

SummaryIn regional hydrologic regression analysis, model selection and validation are regarded as important steps. Here, the model selection is usually based on some measurements of goodness-of-fit between the model prediction and observed data. In Regional Flood Frequency Analysis (RFFA), leave-one-out (LOO) validation or a fixed percentage leave out validation (e.g., 10%) is commonly adopted to assess the predictive ability of regression-based prediction equations. This paper develops a Monte Carlo Cross Validation (MCCV) technique (which has widely been adopted in Chemometrics and Econometrics) in RFFA using Generalised Least Squares Regression (GLSR) and compares it with the most commonly adopted LOO validation approach. The study uses simulated and regional flood data from the state of New South Wales in Australia. It is found that when developing hydrologic regression models, application of the MCCV is likely to result in a more parsimonious model than the LOO. It has also been found that the MCCV can provide a more realistic estimate of a model's predictive ability when compared with the LOO.

Partial Least Squares with Structured Output for Modelling the Metabolomics Data Obtained from Complex Experimental Designs: A Study into the Y-Block Coding.

PubMed

Xu, Yun; Muhamadali, Howbeer; Sayqal, Ali; Dixon, Neil; Goodacre, Royston

2016-10-28

Partial least squares (PLS) is one of the most commonly used supervised modelling approaches for analysing multivariate metabolomics data. PLS is typically employed as either a regression model (PLS-R) or a classification model (PLS-DA). However, in metabolomics studies it is common to investigate multiple, potentially interacting, factors simultaneously following a specific experimental design. Such data often cannot be considered as a "pure" regression or a classification problem. Nevertheless, these data have often still been treated as a regression or classification problem and this could lead to ambiguous results. In this study, we investigated the feasibility of designing a hybrid target matrix Y that better reflects the experimental design than simple regression or binary class membership coding commonly used in PLS modelling. The new design of Y coding was based on the same principle used by structural modelling in machine learning techniques. Two real metabolomics datasets were used as examples to illustrate how the new Y coding can improve the interpretability of the PLS model compared to classic regression/classification coding.

Hierarchical cluster-based partial least squares regression (HC-PLSR) is an efficient tool for metamodelling of nonlinear dynamic models.

PubMed

Tøndel, Kristin; Indahl, Ulf G; Gjuvsland, Arne B; Vik, Jon Olav; Hunter, Peter; Omholt, Stig W; Martens, Harald

2011-06-01

Deterministic dynamic models of complex biological systems contain a large number of parameters and state variables, related through nonlinear differential equations with various types of feedback. A metamodel of such a dynamic model is a statistical approximation model that maps variation in parameters and initial conditions (inputs) to variation in features of the trajectories of the state variables (outputs) throughout the entire biologically relevant input space. A sufficiently accurate mapping can be exploited both instrumentally and epistemically. Multivariate regression methodology is a commonly used approach for emulating dynamic models. However, when the input-output relations are highly nonlinear or non-monotone, a standard linear regression approach is prone to give suboptimal results. We therefore hypothesised that a more accurate mapping can be obtained by locally linear or locally polynomial regression. We present here a new method for local regression modelling, Hierarchical Cluster-based PLS regression (HC-PLSR), where fuzzy C-means clustering is used to separate the data set into parts according to the structure of the response surface. We compare the metamodelling performance of HC-PLSR with polynomial partial least squares regression (PLSR) and ordinary least squares (OLS) regression on various systems: six different gene regulatory network models with various types of feedback, a deterministic mathematical model of the mammalian circadian clock and a model of the mouse ventricular myocyte function. Our results indicate that multivariate regression is well suited for emulating dynamic models in systems biology. The hierarchical approach turned out to be superior to both polynomial PLSR and OLS regression in all three test cases. The advantage, in terms of explained variance and prediction accuracy, was largest in systems with highly nonlinear functional relationships and in systems with positive feedback loops. HC-PLSR is a promising approach for metamodelling in systems biology, especially for highly nonlinear or non-monotone parameter to phenotype maps. The algorithm can be flexibly adjusted to suit the complexity of the dynamic model behaviour, inviting automation in the metamodelling of complex systems.

Hierarchical Cluster-based Partial Least Squares Regression (HC-PLSR) is an efficient tool for metamodelling of nonlinear dynamic models

PubMed Central

2011-01-01

Background Deterministic dynamic models of complex biological systems contain a large number of parameters and state variables, related through nonlinear differential equations with various types of feedback. A metamodel of such a dynamic model is a statistical approximation model that maps variation in parameters and initial conditions (inputs) to variation in features of the trajectories of the state variables (outputs) throughout the entire biologically relevant input space. A sufficiently accurate mapping can be exploited both instrumentally and epistemically. Multivariate regression methodology is a commonly used approach for emulating dynamic models. However, when the input-output relations are highly nonlinear or non-monotone, a standard linear regression approach is prone to give suboptimal results. We therefore hypothesised that a more accurate mapping can be obtained by locally linear or locally polynomial regression. We present here a new method for local regression modelling, Hierarchical Cluster-based PLS regression (HC-PLSR), where fuzzy C-means clustering is used to separate the data set into parts according to the structure of the response surface. We compare the metamodelling performance of HC-PLSR with polynomial partial least squares regression (PLSR) and ordinary least squares (OLS) regression on various systems: six different gene regulatory network models with various types of feedback, a deterministic mathematical model of the mammalian circadian clock and a model of the mouse ventricular myocyte function. Results Our results indicate that multivariate regression is well suited for emulating dynamic models in systems biology. The hierarchical approach turned out to be superior to both polynomial PLSR and OLS regression in all three test cases. The advantage, in terms of explained variance and prediction accuracy, was largest in systems with highly nonlinear functional relationships and in systems with positive feedback loops. Conclusions HC-PLSR is a promising approach for metamodelling in systems biology, especially for highly nonlinear or non-monotone parameter to phenotype maps. The algorithm can be flexibly adjusted to suit the complexity of the dynamic model behaviour, inviting automation in the metamodelling of complex systems. PMID:21627852

A Generalized Least Squares Regression Approach for Computing Effect Sizes in Single-Case Research: Application Examples

ERIC Educational Resources Information Center

Maggin, Daniel M.; Swaminathan, Hariharan; Rogers, Helen J.; O'Keeffe, Breda V.; Sugai, George; Horner, Robert H.

2011-01-01

A new method for deriving effect sizes from single-case designs is proposed. The strategy is applicable to small-sample time-series data with autoregressive errors. The method uses Generalized Least Squares (GLS) to model the autocorrelation of the data and estimate regression parameters to produce an effect size that represents the magnitude of…

Parameter estimation of Monod model by the Least-Squares method for microalgae Botryococcus Braunii sp

NASA Astrophysics Data System (ADS)

See, J. J.; Jamaian, S. S.; Salleh, R. M.; Nor, M. E.; Aman, F.

2018-04-01

This research aims to estimate the parameters of Monod model of microalgae Botryococcus Braunii sp growth by the Least-Squares method. Monod equation is a non-linear equation which can be transformed into a linear equation form and it is solved by implementing the Least-Squares linear regression method. Meanwhile, Gauss-Newton method is an alternative method to solve the non-linear Least-Squares problem with the aim to obtain the parameters value of Monod model by minimizing the sum of square error ( SSE). As the result, the parameters of the Monod model for microalgae Botryococcus Braunii sp can be estimated by the Least-Squares method. However, the estimated parameters value obtained by the non-linear Least-Squares method are more accurate compared to the linear Least-Squares method since the SSE of the non-linear Least-Squares method is less than the linear Least-Squares method.

Mixed geographically weighted regression (MGWR) model with weighted adaptive bi-square for case of dengue hemorrhagic fever (DHF) in Surakarta

NASA Astrophysics Data System (ADS)

Astuti, H. N.; Saputro, D. R. S.; Susanti, Y.

2017-06-01

MGWR model is combination of linear regression model and geographically weighted regression (GWR) model, therefore, MGWR model could produce parameter estimation that had global parameter estimation, and other parameter that had local parameter in accordance with its observation location. The linkage between locations of the observations expressed in specific weighting that is adaptive bi-square. In this research, we applied MGWR model with weighted adaptive bi-square for case of DHF in Surakarta based on 10 factors (variables) that is supposed to influence the number of people with DHF. The observation unit in the research is 51 urban villages and the variables are number of inhabitants, number of houses, house index, many public places, number of healthy homes, number of Posyandu, area width, level population density, welfare of the family, and high-region. Based on this research, we obtained 51 MGWR models. The MGWR model were divided into 4 groups with significant variable is house index as a global variable, an area width as a local variable and the remaining variables vary in each. Global variables are variables that significantly affect all locations, while local variables are variables that significantly affect a specific location.

Electricity Consumption in the Industrial Sector of Jordan: Application of Multivariate Linear Regression and Adaptive Neuro-Fuzzy Techniques

NASA Astrophysics Data System (ADS)

Samhouri, M.; Al-Ghandoor, A.; Fouad, R. H.

2009-08-01

In this study two techniques, for modeling electricity consumption of the Jordanian industrial sector, are presented: (i) multivariate linear regression and (ii) neuro-fuzzy models. Electricity consumption is modeled as function of different variables such as number of establishments, number of employees, electricity tariff, prevailing fuel prices, production outputs, capacity utilizations, and structural effects. It was found that industrial production and capacity utilization are the most important variables that have significant effect on future electrical power demand. The results showed that both the multivariate linear regression and neuro-fuzzy models are generally comparable and can be used adequately to simulate industrial electricity consumption. However, comparison that is based on the square root average squared error of data suggests that the neuro-fuzzy model performs slightly better for future prediction of electricity consumption than the multivariate linear regression model. Such results are in full agreement with similar work, using different methods, for other countries.

The Multivariate Regression Statistics Strategy to Investigate Content-Effect Correlation of Multiple Components in Traditional Chinese Medicine Based on a Partial Least Squares Method.

PubMed

Peng, Ying; Li, Su-Ning; Pei, Xuexue; Hao, Kun

2018-03-01

Amultivariate regression statisticstrategy was developed to clarify multi-components content-effect correlation ofpanaxginseng saponins extract and predict the pharmacological effect by components content. In example 1, firstly, we compared pharmacological effects between panax ginseng saponins extract and individual saponin combinations. Secondly, we examined the anti-platelet aggregation effect in seven different saponin combinations of ginsenoside Rb1, Rg1, Rh, Rd, Ra3 and notoginsenoside R1. Finally, the correlation between anti-platelet aggregation and the content of multiple components was analyzed by a partial least squares algorithm. In example 2, firstly, 18 common peaks were identified in ten different batches of panax ginseng saponins extracts from different origins. Then, we investigated the anti-myocardial ischemia reperfusion injury effects of the ten different panax ginseng saponins extracts. Finally, the correlation between the fingerprints and the cardioprotective effects was analyzed by a partial least squares algorithm. Both in example 1 and 2, the relationship between the components content and pharmacological effect was modeled well by the partial least squares regression equations. Importantly, the predicted effect curve was close to the observed data of dot marked on the partial least squares regression model. This study has given evidences that themulti-component content is a promising information for predicting the pharmacological effects of traditional Chinese medicine.

Impact of multicollinearity on small sample hydrologic regression models

NASA Astrophysics Data System (ADS)

Kroll, Charles N.; Song, Peter

2013-06-01

Often hydrologic regression models are developed with ordinary least squares (OLS) procedures. The use of OLS with highly correlated explanatory variables produces multicollinearity, which creates highly sensitive parameter estimators with inflated variances and improper model selection. It is not clear how to best address multicollinearity in hydrologic regression models. Here a Monte Carlo simulation is developed to compare four techniques to address multicollinearity: OLS, OLS with variance inflation factor screening (VIF), principal component regression (PCR), and partial least squares regression (PLS). The performance of these four techniques was observed for varying sample sizes, correlation coefficients between the explanatory variables, and model error variances consistent with hydrologic regional regression models. The negative effects of multicollinearity are magnified at smaller sample sizes, higher correlations between the variables, and larger model error variances (smaller R2). The Monte Carlo simulation indicates that if the true model is known, multicollinearity is present, and the estimation and statistical testing of regression parameters are of interest, then PCR or PLS should be employed. If the model is unknown, or if the interest is solely on model predictions, is it recommended that OLS be employed since using more complicated techniques did not produce any improvement in model performance. A leave-one-out cross-validation case study was also performed using low-streamflow data sets from the eastern United States. Results indicate that OLS with stepwise selection generally produces models across study regions with varying levels of multicollinearity that are as good as biased regression techniques such as PCR and PLS.

A New Global Regression Analysis Method for the Prediction of Wind Tunnel Model Weight Corrections

NASA Technical Reports Server (NTRS)

Ulbrich, Norbert Manfred; Bridge, Thomas M.; Amaya, Max A.

2014-01-01

A new global regression analysis method is discussed that predicts wind tunnel model weight corrections for strain-gage balance loads during a wind tunnel test. The method determines corrections by combining "wind-on" model attitude measurements with least squares estimates of the model weight and center of gravity coordinates that are obtained from "wind-off" data points. The method treats the least squares fit of the model weight separate from the fit of the center of gravity coordinates. Therefore, it performs two fits of "wind- off" data points and uses the least squares estimator of the model weight as an input for the fit of the center of gravity coordinates. Explicit equations for the least squares estimators of the weight and center of gravity coordinates are derived that simplify the implementation of the method in the data system software of a wind tunnel. In addition, recommendations for sets of "wind-off" data points are made that take typical model support system constraints into account. Explicit equations of the confidence intervals on the model weight and center of gravity coordinates and two different error analyses of the model weight prediction are also discussed in the appendices of the paper.

Comparison of Logistic Regression and Artificial Neural Network in Low Back Pain Prediction: Second National Health Survey

PubMed Central

Parsaeian, M; Mohammad, K; Mahmoudi, M; Zeraati, H

2012-01-01

Background: The purpose of this investigation was to compare empirically predictive ability of an artificial neural network with a logistic regression in prediction of low back pain. Methods: Data from the second national health survey were considered in this investigation. This data includes the information of low back pain and its associated risk factors among Iranian people aged 15 years and older. Artificial neural network and logistic regression models were developed using a set of 17294 data and they were validated in a test set of 17295 data. Hosmer and Lemeshow recommendation for model selection was used in fitting the logistic regression. A three-layer perceptron with 9 inputs, 3 hidden and 1 output neurons was employed. The efficiency of two models was compared by receiver operating characteristic analysis, root mean square and -2 Loglikelihood criteria. Results: The area under the ROC curve (SE), root mean square and -2Loglikelihood of the logistic regression was 0.752 (0.004), 0.3832 and 14769.2, respectively. The area under the ROC curve (SE), root mean square and -2Loglikelihood of the artificial neural network was 0.754 (0.004), 0.3770 and 14757.6, respectively. Conclusions: Based on these three criteria, artificial neural network would give better performance than logistic regression. Although, the difference is statistically significant, it does not seem to be clinically significant. PMID:23113198

Comparison of logistic regression and artificial neural network in low back pain prediction: second national health survey.

PubMed

Parsaeian, M; Mohammad, K; Mahmoudi, M; Zeraati, H

2012-01-01

The purpose of this investigation was to compare empirically predictive ability of an artificial neural network with a logistic regression in prediction of low back pain. Data from the second national health survey were considered in this investigation. This data includes the information of low back pain and its associated risk factors among Iranian people aged 15 years and older. Artificial neural network and logistic regression models were developed using a set of 17294 data and they were validated in a test set of 17295 data. Hosmer and Lemeshow recommendation for model selection was used in fitting the logistic regression. A three-layer perceptron with 9 inputs, 3 hidden and 1 output neurons was employed. The efficiency of two models was compared by receiver operating characteristic analysis, root mean square and -2 Loglikelihood criteria. The area under the ROC curve (SE), root mean square and -2Loglikelihood of the logistic regression was 0.752 (0.004), 0.3832 and 14769.2, respectively. The area under the ROC curve (SE), root mean square and -2Loglikelihood of the artificial neural network was 0.754 (0.004), 0.3770 and 14757.6, respectively. Based on these three criteria, artificial neural network would give better performance than logistic regression. Although, the difference is statistically significant, it does not seem to be clinically significant.

On Insensitivity of the Chi-Square Model Test to Nonlinear Misspecification in Structural Equation Models

ERIC Educational Resources Information Center

Mooijaart, Ab; Satorra, Albert

2009-01-01

In this paper, we show that for some structural equation models (SEM), the classical chi-square goodness-of-fit test is unable to detect the presence of nonlinear terms in the model. As an example, we consider a regression model with latent variables and interactions terms. Not only the model test has zero power against that type of…

Ridge: a computer program for calculating ridge regression estimates

Treesearch

Donald E. Hilt; Donald W. Seegrist

1977-01-01

Least-squares coefficients for multiple-regression models may be unstable when the independent variables are highly correlated. Ridge regression is a biased estimation procedure that produces stable estimates of the coefficients. Ridge regression is discussed, and a computer program for calculating the ridge coefficients is presented.

Extension of the Haseman-Elston regression model to longitudinal data.

PubMed

Won, Sungho; Elston, Robert C; Park, Taesung

2006-01-01

We propose an extension to longitudinal data of the Haseman and Elston regression method for linkage analysis. The proposed model is a mixed model having several random effects. As response variable, we investigate the sibship sample mean corrected cross-product (smHE) and the BLUP-mean corrected cross product (pmHE), comparing them with the original squared difference (oHE), the overall mean corrected cross-product (rHE), and the weighted average of the squared difference and the squared mean-corrected sum (wHE). The proposed model allows for the correlation structure of longitudinal data. Also, the model can test for gene x time interaction to discover genetic variation over time. The model was applied in an analysis of the Genetic Analysis Workshop 13 (GAW13) simulated dataset for a quantitative trait simulating systolic blood pressure. Independence models did not preserve the test sizes, while the mixed models with both family and sibpair random effects tended to preserve size well. Copyright 2006 S. Karger AG, Basel.

Evaluation of three statistical prediction models for forensic age prediction based on DNA methylation.

PubMed

Smeers, Inge; Decorte, Ronny; Van de Voorde, Wim; Bekaert, Bram

2018-05-01

DNA methylation is a promising biomarker for forensic age prediction. A challenge that has emerged in recent studies is the fact that prediction errors become larger with increasing age due to interindividual differences in epigenetic ageing rates. This phenomenon of non-constant variance or heteroscedasticity violates an assumption of the often used method of ordinary least squares (OLS) regression. The aim of this study was to evaluate alternative statistical methods that do take heteroscedasticity into account in order to provide more accurate, age-dependent prediction intervals. A weighted least squares (WLS) regression is proposed as well as a quantile regression model. Their performances were compared against an OLS regression model based on the same dataset. Both models provided age-dependent prediction intervals which account for the increasing variance with age, but WLS regression performed better in terms of success rate in the current dataset. However, quantile regression might be a preferred method when dealing with a variance that is not only non-constant, but also not normally distributed. Ultimately the choice of which model to use should depend on the observed characteristics of the data. Copyright © 2018 Elsevier B.V. All rights reserved.

Membrane Introduction Mass Spectrometry Combined with an Orthogonal Partial-Least Squares Calibration Model for Mixture Analysis.

PubMed

Li, Min; Zhang, Lu; Yao, Xiaolong; Jiang, Xingyu

2017-01-01

The emerging membrane introduction mass spectrometry technique has been successfully used to detect benzene, toluene, ethyl benzene and xylene (BTEX), while overlapped spectra have unfortunately hindered its further application to the analysis of mixtures. Multivariate calibration, an efficient method to analyze mixtures, has been widely applied. In this paper, we compared univariate and multivariate analyses for quantification of the individual components of mixture samples. The results showed that the univariate analysis creates poor models with regression coefficients of 0.912, 0.867, 0.440 and 0.351 for BTEX, respectively. For multivariate analysis, a comparison to the partial-least squares (PLS) model shows that the orthogonal partial-least squares (OPLS) regression exhibits an optimal performance with regression coefficients of 0.995, 0.999, 0.980 and 0.976, favorable calibration parameters (RMSEC and RMSECV) and a favorable validation parameter (RMSEP). Furthermore, the OPLS exhibits a good recovery of 73.86 - 122.20% and relative standard deviation (RSD) of the repeatability of 1.14 - 4.87%. Thus, MIMS coupled with the OPLS regression provides an optimal approach for a quantitative BTEX mixture analysis in monitoring and predicting water pollution.

Least-Squares Regression and Spectral Residual Augmented Classical Least-Squares Chemometric Models for Stability-Indicating Analysis of Agomelatine and Its Degradation Products: A Comparative Study.

PubMed

Naguib, Ibrahim A; Abdelrahman, Maha M; El Ghobashy, Mohamed R; Ali, Nesma A

2016-01-01

Two accurate, sensitive, and selective stability-indicating methods are developed and validated for simultaneous quantitative determination of agomelatine (AGM) and its forced degradation products (Deg I and Deg II), whether in pure forms or in pharmaceutical formulations. Partial least-squares regression (PLSR) and spectral residual augmented classical least-squares (SRACLS) are two chemometric models that are being subjected to a comparative study through handling UV spectral data in range (215-350 nm). For proper analysis, a three-factor, four-level experimental design was established, resulting in a training set consisting of 16 mixtures containing different ratios of interfering species. An independent test set consisting of eight mixtures was used to validate the prediction ability of the suggested models. The results presented indicate the ability of mentioned multivariate calibration models to analyze AGM, Deg I, and Deg II with high selectivity and accuracy. The analysis results of the pharmaceutical formulations were statistically compared to the reference HPLC method, with no significant differences observed regarding accuracy and precision. The SRACLS model gives comparable results to the PLSR model; however, it keeps the qualitative spectral information of the classical least-squares algorithm for analyzed components.

The Collinearity Free and Bias Reduced Regression Estimation Project: The Theory of Normalization Ridge Regression. Report No. 2.

ERIC Educational Resources Information Center

Bulcock, J. W.; And Others

Multicollinearity refers to the presence of highly intercorrelated independent variables in structural equation models, that is, models estimated by using techniques such as least squares regression and maximum likelihood. There is a problem of multicollinearity in both the natural and social sciences where theory formulation and estimation is in…

Analysis of Learning Curve Fitting Techniques.

DTIC Science & Technology

1987-09-01

1986. 15. Neter, John and others. Applied Linear Regression Models. Homewood IL: Irwin, 19-33. 16. SAS User’s Guide: Basics, Version 5 Edition. SAS... Linear Regression Techniques (15:23-52). Random errors are assumed to be normally distributed when using -# ordinary least-squares, according to Johnston...lot estimated by the improvement curve formula. For a more detailed explanation of the ordinary least-squares technique, see Neter, et. al., Applied

Partial least squares for efficient models of fecal indicator bacteria on Great Lakes beaches

USGS Publications Warehouse

Brooks, Wesley R.; Fienen, Michael N.; Corsi, Steven R.

2013-01-01

At public beaches, it is now common to mitigate the impact of water-borne pathogens by posting a swimmer's advisory when the concentration of fecal indicator bacteria (FIB) exceeds an action threshold. Since culturing the bacteria delays public notification when dangerous conditions exist, regression models are sometimes used to predict the FIB concentration based on readily-available environmental measurements. It is hard to know which environmental parameters are relevant to predicting FIB concentration, and the parameters are usually correlated, which can hurt the predictive power of a regression model. Here the method of partial least squares (PLS) is introduced to automate the regression modeling process. Model selection is reduced to the process of setting a tuning parameter to control the decision threshold that separates predicted exceedances of the standard from predicted non-exceedances. The method is validated by application to four Great Lakes beaches during the summer of 2010. Performance of the PLS models compares favorably to that of the existing state-of-the-art regression models at these four sites.

Estimation of Ordinary Differential Equation Parameters Using Constrained Local Polynomial Regression.

PubMed

Ding, A Adam; Wu, Hulin

2014-10-01

We propose a new method to use a constrained local polynomial regression to estimate the unknown parameters in ordinary differential equation models with a goal of improving the smoothing-based two-stage pseudo-least squares estimate. The equation constraints are derived from the differential equation model and are incorporated into the local polynomial regression in order to estimate the unknown parameters in the differential equation model. We also derive the asymptotic bias and variance of the proposed estimator. Our simulation studies show that our new estimator is clearly better than the pseudo-least squares estimator in estimation accuracy with a small price of computational cost. An application example on immune cell kinetics and trafficking for influenza infection further illustrates the benefits of the proposed new method.

Estimation of Ordinary Differential Equation Parameters Using Constrained Local Polynomial Regression

PubMed Central

Ding, A. Adam; Wu, Hulin

2015-01-01

We propose a new method to use a constrained local polynomial regression to estimate the unknown parameters in ordinary differential equation models with a goal of improving the smoothing-based two-stage pseudo-least squares estimate. The equation constraints are derived from the differential equation model and are incorporated into the local polynomial regression in order to estimate the unknown parameters in the differential equation model. We also derive the asymptotic bias and variance of the proposed estimator. Our simulation studies show that our new estimator is clearly better than the pseudo-least squares estimator in estimation accuracy with a small price of computational cost. An application example on immune cell kinetics and trafficking for influenza infection further illustrates the benefits of the proposed new method. PMID:26401093

Weighted regression analysis and interval estimators

Treesearch

Donald W. Seegrist

1974-01-01

A method for deriving the weighted least squares estimators for the parameters of a multiple regression model. Confidence intervals for expected values, and prediction intervals for the means of future samples are given.

Application of nonlinear least-squares regression to ground-water flow modeling, west-central Florida

USGS Publications Warehouse

Yobbi, D.K.

2000-01-01

A nonlinear least-squares regression technique for estimation of ground-water flow model parameters was applied to an existing model of the regional aquifer system underlying west-central Florida. The regression technique minimizes the differences between measured and simulated water levels. Regression statistics, including parameter sensitivities and correlations, were calculated for reported parameter values in the existing model. Optimal parameter values for selected hydrologic variables of interest are estimated by nonlinear regression. Optimal estimates of parameter values are about 140 times greater than and about 0.01 times less than reported values. Independently estimating all parameters by nonlinear regression was impossible, given the existing zonation structure and number of observations, because of parameter insensitivity and correlation. Although the model yields parameter values similar to those estimated by other methods and reproduces the measured water levels reasonably accurately, a simpler parameter structure should be considered. Some possible ways of improving model calibration are to: (1) modify the defined parameter-zonation structure by omitting and/or combining parameters to be estimated; (2) carefully eliminate observation data based on evidence that they are likely to be biased; (3) collect additional water-level data; (4) assign values to insensitive parameters, and (5) estimate the most sensitive parameters first, then, using the optimized values for these parameters, estimate the entire data set.

Error Covariance Penalized Regression: A novel multivariate model combining penalized regression with multivariate error structure.

PubMed

Allegrini, Franco; Braga, Jez W B; Moreira, Alessandro C O; Olivieri, Alejandro C

2018-06-29

A new multivariate regression model, named Error Covariance Penalized Regression (ECPR) is presented. Following a penalized regression strategy, the proposed model incorporates information about the measurement error structure of the system, using the error covariance matrix (ECM) as a penalization term. Results are reported from both simulations and experimental data based on replicate mid and near infrared (MIR and NIR) spectral measurements. The results for ECPR are better under non-iid conditions when compared with traditional first-order multivariate methods such as ridge regression (RR), principal component regression (PCR) and partial least-squares regression (PLS). Copyright © 2018 Elsevier B.V. All rights reserved.

Analyzing Multilevel Data: An Empirical Comparison of Parameter Estimates of Hierarchical Linear Modeling and Ordinary Least Squares Regression

ERIC Educational Resources Information Center

Rocconi, Louis M.

2011-01-01

Hierarchical linear models (HLM) solve the problems associated with the unit of analysis problem such as misestimated standard errors, heterogeneity of regression and aggregation bias by modeling all levels of interest simultaneously. Hierarchical linear modeling resolves the problem of misestimated standard errors by incorporating a unique random…

A weighted least squares estimation of the polynomial regression model on paddy production in the area of Kedah and Perlis

NASA Astrophysics Data System (ADS)

Musa, Rosliza; Ali, Zalila; Baharum, Adam; Nor, Norlida Mohd

2017-08-01

The linear regression model assumes that all random error components are identically and independently distributed with constant variance. Hence, each data point provides equally precise information about the deterministic part of the total variation. In other words, the standard deviations of the error terms are constant over all values of the predictor variables. When the assumption of constant variance is violated, the ordinary least squares estimator of regression coefficient lost its property of minimum variance in the class of linear and unbiased estimators. Weighted least squares estimation are often used to maximize the efficiency of parameter estimation. A procedure that treats all of the data equally would give less precisely measured points more influence than they should have and would give highly precise points too little influence. Optimizing the weighted fitting criterion to find the parameter estimates allows the weights to determine the contribution of each observation to the final parameter estimates. This study used polynomial model with weighted least squares estimation to investigate paddy production of different paddy lots based on paddy cultivation characteristics and environmental characteristics in the area of Kedah and Perlis. The results indicated that factors affecting paddy production are mixture fertilizer application cycle, average temperature, the squared effect of average rainfall, the squared effect of pest and disease, the interaction between acreage with amount of mixture fertilizer, the interaction between paddy variety and NPK fertilizer application cycle and the interaction between pest and disease and NPK fertilizer application cycle.

Partial Least Squares Regression Models for the Analysis of Kinase Signaling.

PubMed

Bourgeois, Danielle L; Kreeger, Pamela K

2017-01-01

Partial least squares regression (PLSR) is a data-driven modeling approach that can be used to analyze multivariate relationships between kinase networks and cellular decisions or patient outcomes. In PLSR, a linear model relating an X matrix of dependent variables and a Y matrix of independent variables is generated by extracting the factors with the strongest covariation. While the identified relationship is correlative, PLSR models can be used to generate quantitative predictions for new conditions or perturbations to the network, allowing for mechanisms to be identified. This chapter will provide a brief explanation of PLSR and provide an instructive example to demonstrate the use of PLSR to analyze kinase signaling.

Modified locally weighted--partial least squares regression improving clinical predictions from infrared spectra of human serum samples.

PubMed

Perez-Guaita, David; Kuligowski, Julia; Quintás, Guillermo; Garrigues, Salvador; Guardia, Miguel de la

2013-03-30

Locally weighted partial least squares regression (LW-PLSR) has been applied to the determination of four clinical parameters in human serum samples (total protein, triglyceride, glucose and urea contents) by Fourier transform infrared (FTIR) spectroscopy. Classical LW-PLSR models were constructed using different spectral regions. For the selection of parameters by LW-PLSR modeling, a multi-parametric study was carried out employing the minimum root-mean square error of cross validation (RMSCV) as objective function. In order to overcome the effect of strong matrix interferences on the predictive accuracy of LW-PLSR models, this work focuses on sample selection. Accordingly, a novel strategy for the development of local models is proposed. It was based on the use of: (i) principal component analysis (PCA) performed on an analyte specific spectral region for identifying most similar sample spectra and (ii) partial least squares regression (PLSR) constructed using the whole spectrum. Results found by using this strategy were compared to those provided by PLSR using the same spectral intervals as for LW-PLSR. Prediction errors found by both, classical and modified LW-PLSR improved those obtained by PLSR. Hence, both proposed approaches were useful for the determination of analytes present in a complex matrix as in the case of human serum samples. Copyright © 2013 Elsevier B.V. All rights reserved.

Applying Regression Analysis to Problems in Institutional Research.

ERIC Educational Resources Information Center

Bohannon, Tom R.

1988-01-01

Regression analysis is one of the most frequently used statistical techniques in institutional research. Principles of least squares, model building, residual analysis, influence statistics, and multi-collinearity are described and illustrated. (Author/MSE)

Functional Relationships and Regression Analysis.

ERIC Educational Resources Information Center

Preece, Peter F. W.

1978-01-01

Using a degenerate multivariate normal model for the distribution of organismic variables, the form of least-squares regression analysis required to estimate a linear functional relationship between variables is derived. It is suggested that the two conventional regression lines may be considered to describe functional, not merely statistical,…

Intrinsic Raman spectroscopy for quantitative biological spectroscopy Part II

PubMed Central

Bechtel, Kate L.; Shih, Wei-Chuan; Feld, Michael S.

2009-01-01

We demonstrate the effectiveness of intrinsic Raman spectroscopy (IRS) at reducing errors caused by absorption and scattering. Physical tissue models, solutions of varying absorption and scattering coefficients with known concentrations of Raman scatterers, are studied. We show significant improvement in prediction error by implementing IRS to predict concentrations of Raman scatterers using both ordinary least squares regression (OLS) and partial least squares regression (PLS). In particular, we show that IRS provides a robust calibration model that does not increase in error when applied to samples with optical properties outside the range of calibration. PMID:18711512

Least Square Regression Method for Estimating Gas Concentration in an Electronic Nose System

PubMed Central

Khalaf, Walaa; Pace, Calogero; Gaudioso, Manlio

2009-01-01

We describe an Electronic Nose (ENose) system which is able to identify the type of analyte and to estimate its concentration. The system consists of seven sensors, five of them being gas sensors (supplied with different heater voltage values), the remainder being a temperature and a humidity sensor, respectively. To identify a new analyte sample and then to estimate its concentration, we use both some machine learning techniques and the least square regression principle. In fact, we apply two different training models; the first one is based on the Support Vector Machine (SVM) approach and is aimed at teaching the system how to discriminate among different gases, while the second one uses the least squares regression approach to predict the concentration of each type of analyte. PMID:22573980

Development of a partial least squares-artificial neural network (PLS-ANN) hybrid model for the prediction of consumer liking scores of ready-to-drink green tea beverages.

PubMed

Yu, Peigen; Low, Mei Yin; Zhou, Weibiao

2018-01-01

In order to develop products that would be preferred by consumers, the effects of the chemical compositions of ready-to-drink green tea beverages on consumer liking were studied through regression analyses. Green tea model systems were prepared by dosing solutions of 0.1% green tea extract with differing concentrations of eight flavour keys deemed to be important for green tea aroma and taste, based on a D-optimal experimental design, before undergoing commercial sterilisation. Sensory evaluation of the green tea model system was carried out using an untrained consumer panel to obtain hedonic liking scores of the samples. Regression models were subsequently trained to objectively predict the consumer liking scores of the green tea model systems. A linear partial least squares (PLS) regression model was developed to describe the effects of the eight flavour keys on consumer liking, with a coefficient of determination (R 2 ) of 0.733, and a root-mean-square error (RMSE) of 3.53%. The PLS model was further augmented with an artificial neural network (ANN) to establish a PLS-ANN hybrid model. The established hybrid model was found to give a better prediction of consumer liking scores, based on its R 2 (0.875) and RMSE (2.41%). Copyright © 2017 Elsevier Ltd. All rights reserved.

The Variance Normalization Method of Ridge Regression Analysis.

ERIC Educational Resources Information Center

Bulcock, J. W.; And Others

The testing of contemporary sociological theory often calls for the application of structural-equation models to data which are inherently collinear. It is shown that simple ridge regression, which is commonly used for controlling the instability of ordinary least squares regression estimates in ill-conditioned data sets, is not a legitimate…

Independent contrasts and PGLS regression estimators are equivalent.

PubMed

Blomberg, Simon P; Lefevre, James G; Wells, Jessie A; Waterhouse, Mary

2012-05-01

We prove that the slope parameter of the ordinary least squares regression of phylogenetically independent contrasts (PICs) conducted through the origin is identical to the slope parameter of the method of generalized least squares (GLSs) regression under a Brownian motion model of evolution. This equivalence has several implications: 1. Understanding the structure of the linear model for GLS regression provides insight into when and why phylogeny is important in comparative studies. 2. The limitations of the PIC regression analysis are the same as the limitations of the GLS model. In particular, phylogenetic covariance applies only to the response variable in the regression and the explanatory variable should be regarded as fixed. Calculation of PICs for explanatory variables should be treated as a mathematical idiosyncrasy of the PIC regression algorithm. 3. Since the GLS estimator is the best linear unbiased estimator (BLUE), the slope parameter estimated using PICs is also BLUE. 4. If the slope is estimated using different branch lengths for the explanatory and response variables in the PIC algorithm, the estimator is no longer the BLUE, so this is not recommended. Finally, we discuss whether or not and how to accommodate phylogenetic covariance in regression analyses, particularly in relation to the problem of phylogenetic uncertainty. This discussion is from both frequentist and Bayesian perspectives.

Mitigating Errors in External Respiratory Surrogate-Based Models of Tumor Position

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

Malinowski, Kathleen T.; Fischell Department of Bioengineering, University of Maryland, College Park, MD; McAvoy, Thomas J.

2012-04-01

Purpose: To investigate the effect of tumor site, measurement precision, tumor-surrogate correlation, training data selection, model design, and interpatient and interfraction variations on the accuracy of external marker-based models of tumor position. Methods and Materials: Cyberknife Synchrony system log files comprising synchronously acquired positions of external markers and the tumor from 167 treatment fractions were analyzed. The accuracy of Synchrony, ordinary-least-squares regression, and partial-least-squares regression models for predicting the tumor position from the external markers was evaluated. The quantity and timing of the data used to build the predictive model were varied. The effects of tumor-surrogate correlation and the precisionmore » in both the tumor and the external surrogate position measurements were explored by adding noise to the data. Results: The tumor position prediction errors increased during the duration of a fraction. Increasing the training data quantities did not always lead to more accurate models. Adding uncorrelated noise to the external marker-based inputs degraded the tumor-surrogate correlation models by 16% for partial-least-squares and 57% for ordinary-least-squares. External marker and tumor position measurement errors led to tumor position prediction changes 0.3-3.6 times the magnitude of the measurement errors, varying widely with model algorithm. The tumor position prediction errors were significantly associated with the patient index but not with the fraction index or tumor site. Partial-least-squares was as accurate as Synchrony and more accurate than ordinary-least-squares. Conclusions: The accuracy of surrogate-based inferential models of tumor position was affected by all the investigated factors, except for the tumor site and fraction index.« less

Model Estimation Using Ridge Regression with the Variance Normalization Criterion. Interim Report No. 2. The Education and Inequality in Canada Project.

ERIC Educational Resources Information Center

Lee, Wan-Fung; Bulcock, Jeffrey Wilson

The purposes of this study are: (1) to demonstrate the superiority of simple ridge regression over ordinary least squares regression through theoretical argument and empirical example; (2) to modify ridge regression through use of the variance normalization criterion; and (3) to demonstrate the superiority of simple ridge regression based on the…

Application of least square support vector machine and multivariate adaptive regression spline models in long term prediction of river water pollution

NASA Astrophysics Data System (ADS)

Kisi, Ozgur; Parmar, Kulwinder Singh

2016-03-01

This study investigates the accuracy of least square support vector machine (LSSVM), multivariate adaptive regression splines (MARS) and M5 model tree (M5Tree) in modeling river water pollution. Various combinations of water quality parameters, Free Ammonia (AMM), Total Kjeldahl Nitrogen (TKN), Water Temperature (WT), Total Coliform (TC), Fecal Coliform (FC) and Potential of Hydrogen (pH) monitored at Nizamuddin, Delhi Yamuna River in India were used as inputs to the applied models. Results indicated that the LSSVM and MARS models had almost same accuracy and they performed better than the M5Tree model in modeling monthly chemical oxygen demand (COD). The average root mean square error (RMSE) of the LSSVM and M5Tree models was decreased by 1.47% and 19.1% using MARS model, respectively. Adding TC input to the models did not increase their accuracy in modeling COD while adding FC and pH inputs to the models generally decreased the accuracy. The overall results indicated that the MARS and LSSVM models could be successfully used in estimating monthly river water pollution level by using AMM, TKN and WT parameters as inputs.

Modified Regression Correlation Coefficient for Poisson Regression Model

NASA Astrophysics Data System (ADS)

Kaengthong, Nattacha; Domthong, Uthumporn

2017-09-01

This study gives attention to indicators in predictive power of the Generalized Linear Model (GLM) which are widely used; however, often having some restrictions. We are interested in regression correlation coefficient for a Poisson regression model. This is a measure of predictive power, and defined by the relationship between the dependent variable (Y) and the expected value of the dependent variable given the independent variables [E(Y|X)] for the Poisson regression model. The dependent variable is distributed as Poisson. The purpose of this research was modifying regression correlation coefficient for Poisson regression model. We also compare the proposed modified regression correlation coefficient with the traditional regression correlation coefficient in the case of two or more independent variables, and having multicollinearity in independent variables. The result shows that the proposed regression correlation coefficient is better than the traditional regression correlation coefficient based on Bias and the Root Mean Square Error (RMSE).

Discriminative least squares regression for multiclass classification and feature selection.

PubMed

Xiang, Shiming; Nie, Feiping; Meng, Gaofeng; Pan, Chunhong; Zhang, Changshui

2012-11-01

This paper presents a framework of discriminative least squares regression (LSR) for multiclass classification and feature selection. The core idea is to enlarge the distance between different classes under the conceptual framework of LSR. First, a technique called ε-dragging is introduced to force the regression targets of different classes moving along opposite directions such that the distances between classes can be enlarged. Then, the ε-draggings are integrated into the LSR model for multiclass classification. Our learning framework, referred to as discriminative LSR, has a compact model form, where there is no need to train two-class machines that are independent of each other. With its compact form, this model can be naturally extended for feature selection. This goal is achieved in terms of L2,1 norm of matrix, generating a sparse learning model for feature selection. The model for multiclass classification and its extension for feature selection are finally solved elegantly and efficiently. Experimental evaluation over a range of benchmark datasets indicates the validity of our method.

Bayesian generalized least squares regression with application to log Pearson type 3 regional skew estimation

NASA Astrophysics Data System (ADS)

Reis, D. S.; Stedinger, J. R.; Martins, E. S.

2005-10-01

This paper develops a Bayesian approach to analysis of a generalized least squares (GLS) regression model for regional analyses of hydrologic data. The new approach allows computation of the posterior distributions of the parameters and the model error variance using a quasi-analytic approach. Two regional skew estimation studies illustrate the value of the Bayesian GLS approach for regional statistical analysis of a shape parameter and demonstrate that regional skew models can be relatively precise with effective record lengths in excess of 60 years. With Bayesian GLS the marginal posterior distribution of the model error variance and the corresponding mean and variance of the parameters can be computed directly, thereby providing a simple but important extension of the regional GLS regression procedures popularized by Tasker and Stedinger (1989), which is sensitive to the likely values of the model error variance when it is small relative to the sampling error in the at-site estimator.

A comparison between the use of Cox regression and the use of partial least squares-Cox regression to predict the survival of kidney-transplant patients

NASA Astrophysics Data System (ADS)

Solimun

2017-05-01

The aim of this research is to model survival data from kidney-transplant patients using the partial least squares (PLS)-Cox regression, which can both meet and not meet the no-multicollinearity assumption. The secondary data were obtained from research entitled "Factors affecting the survival of kidney-transplant patients". The research subjects comprised 250 patients. The predictor variables consisted of: age (X1), sex (X2); two categories, prior hemodialysis duration (X3), diabetes (X4); two categories, prior transplantation number (X5), number of blood transfusions (X6), discrepancy score (X7), use of antilymphocyte globulin(ALG) (X8); two categories, while the response variable was patient survival time (in months). Partial least squares regression is a model that connects the predictor variables X and the response variable y and it initially aims to determine the relationship between them. Results of the above analyses suggest that the survival of kidney transplant recipients ranged from 0 to 55 months, with 62% of the patients surviving until they received treatment that lasted for 55 months. The PLS-Cox regression analysis results revealed that patients' age and the use of ALG significantly affected the survival time of patients. The factor of patients' age (X1) in the PLS-Cox regression model merely affected the failure probability by 1.201. This indicates that the probability of dying for elderly patients with a kidney transplant is 1.152 times higher than that for younger patients.

Causal Models with Unmeasured Variables: An Introduction to LISREL.

ERIC Educational Resources Information Center

Wolfle, Lee M.

Whenever one uses ordinary least squares regression, one is making an implicit assumption that all of the independent variables have been measured without error. Such an assumption is obviously unrealistic for most social data. One approach for estimating such regression models is to measure implied coefficients between latent variables for which…

Soil Cd, Cr, Cu, Ni, Pb and Zn sorption and retention models using SVM: Variable selection and competitive model.

PubMed

González Costa, J J; Reigosa, M J; Matías, J M; Covelo, E F

2017-09-01

The aim of this study was to model the sorption and retention of Cd, Cu, Ni, Pb and Zn in soils. To that extent, the sorption and retention of these metals were studied and the soil characterization was performed separately. Multiple stepwise regression was used to produce multivariate models with linear techniques and with support vector machines, all of which included 15 explanatory variables characterizing soils. When the R-squared values are represented, two different groups are noticed. Cr, Cu and Pb sorption and retention show a higher R-squared; the most explanatory variables being humified organic matter, Al oxides and, in some cases, cation-exchange capacity (CEC). The other group of metals (Cd, Ni and Zn) shows a lower R-squared, and clays are the most explanatory variables, including a percentage of vermiculite and slime. In some cases, quartz, plagioclase or hematite percentages also show some explanatory capacity. Support Vector Machine (SVM) regression shows that the different models are not as regular as in multiple regression in terms of number of variables, the regression for nickel adsorption being the one with the highest number of variables in its optimal model. On the other hand, there are cases where the most explanatory variables are the same for two metals, as it happens with Cd and Cr adsorption. A similar adsorption mechanism is thus postulated. These patterns of the introduction of variables in the model allow us to create explainability sequences. Those which are the most similar to the selectivity sequences obtained by Covelo (2005) are Mn oxides in multiple regression and change capacity in SVM. Among all the variables, the only one that is explanatory for all the metals after applying the maximum parsimony principle is the percentage of sand in the retention process. In the competitive model arising from the aforementioned sequences, the most intense competitiveness for the adsorption and retention of different metals appears between Cr and Cd, Cu and Zn in multiple regression; and between Cr and Cd in SVM regression. Copyright © 2017 Elsevier B.V. All rights reserved.

Geodesic least squares regression on information manifolds

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

Verdoolaege, Geert, E-mail: geert.verdoolaege@ugent.be

We present a novel regression method targeted at situations with significant uncertainty on both the dependent and independent variables or with non-Gaussian distribution models. Unlike the classic regression model, the conditional distribution of the response variable suggested by the data need not be the same as the modeled distribution. Instead they are matched by minimizing the Rao geodesic distance between them. This yields a more flexible regression method that is less constrained by the assumptions imposed through the regression model. As an example, we demonstrate the improved resistance of our method against some flawed model assumptions and we apply thismore » to scaling laws in magnetic confinement fusion.« less

Modelling of the batch biosorption system: study on exchange of protons with cell wall-bound mineral ions.

PubMed

Mishra, Vishal

2015-01-01

The interchange of the protons with the cell wall-bound calcium and magnesium ions at the interface of solution/bacterial cell surface in the biosorption system at various concentrations of protons has been studied in the present work. A mathematical model for establishing the correlation between concentration of protons and active sites was developed and optimized. The sporadic limited residence time reactor was used to titrate the calcium and magnesium ions at the individual data point. The accuracy of the proposed mathematical model was estimated using error functions such as nonlinear regression, adjusted nonlinear regression coefficient, the chi-square test, P-test and F-test. The values of the chi-square test (0.042-0.017), P-test (<0.001-0.04), sum of square errors (0.061-0.016), root mean square error (0.01-0.04) and F-test (2.22-19.92) reported in the present research indicated the suitability of the model over a wide range of proton concentrations. The zeta potential of the bacterium surface at various concentrations of protons was observed to validate the denaturation of active sites.

Fast determination of total ginsenosides content in ginseng powder by near infrared reflectance spectroscopy

NASA Astrophysics Data System (ADS)

Chen, Hua-cai; Chen, Xing-dan; Lu, Yong-jun; Cao, Zhi-qiang

2006-01-01

Near infrared (NIR) reflectance spectroscopy was used to develop a fast determination method for total ginsenosides in Ginseng (Panax Ginseng) powder. The spectra were analyzed with multiplicative signal correction (MSC) correlation method. The best correlative spectra region with the total ginsenosides content was 1660 nm~1880 nm and 2230nm~2380 nm. The NIR calibration models of ginsenosides were built with multiple linear regression (MLR), principle component regression (PCR) and partial least squares (PLS) regression respectively. The results showed that the calibration model built with PLS combined with MSC and the optimal spectrum region was the best one. The correlation coefficient and the root mean square error of correction validation (RMSEC) of the best calibration model were 0.98 and 0.15% respectively. The optimal spectrum region for calibration was 1204nm~2014nm. The result suggested that using NIR to rapidly determinate the total ginsenosides content in ginseng powder were feasible.

Comparing The Effectiveness of a90/95 Calculations (Preprint)

DTIC Science & Technology

2006-09-01

Nachtsheim, John Neter, William Li, Applied Linear Statistical Models , 5th ed., McGraw-Hill/Irwin, 2005 5. Mood, Graybill and Boes, Introduction...curves is based on methods that are only valid for ordinary linear regression. Requirements for a valid Ordinary Least-Squares Regression Model There... linear . For example is a linear model ; is not. 2. Uniform variance (homoscedasticity

A regression-kriging model for estimation of rainfall in the Laohahe basin

NASA Astrophysics Data System (ADS)

Wang, Hong; Ren, Li L.; Liu, Gao H.

2009-10-01

This paper presents a multivariate geostatistical algorithm called regression-kriging (RK) for predicting the spatial distribution of rainfall by incorporating five topographic/geographic factors of latitude, longitude, altitude, slope and aspect. The technique is illustrated using rainfall data collected at 52 rain gauges from the Laohahe basis in northeast China during 1986-2005 . Rainfall data from 44 stations were selected for modeling and the remaining 8 stations were used for model validation. To eliminate multicollinearity, the five explanatory factors were first transformed using factor analysis with three Principal Components (PCs) extracted. The rainfall data were then fitted using step-wise regression and residuals interpolated using SK. The regression coefficients were estimated by generalized least squares (GLS), which takes the spatial heteroskedasticity between rainfall and PCs into account. Finally, the rainfall prediction based on RK was compared with that predicted from ordinary kriging (OK) and ordinary least squares (OLS) multiple regression (MR). For correlated topographic factors are taken into account, RK improves the efficiency of predictions. RK achieved a lower relative root mean square error (RMSE) (44.67%) than MR (49.23%) and OK (73.60%) and a lower bias than MR and OK (23.82 versus 30.89 and 32.15 mm) for annual rainfall. It is much more effective for the wet season than for the dry season. RK is suitable for estimation of rainfall in areas where there are no stations nearby and where topography has a major influence on rainfall.

Analysis of Sting Balance Calibration Data Using Optimized Regression Models

NASA Technical Reports Server (NTRS)

Ulbrich, Norbert; Bader, Jon B.

2009-01-01

Calibration data of a wind tunnel sting balance was processed using a search algorithm that identifies an optimized regression model for the data analysis. The selected sting balance had two moment gages that were mounted forward and aft of the balance moment center. The difference and the sum of the two gage outputs were fitted in the least squares sense using the normal force and the pitching moment at the balance moment center as independent variables. The regression model search algorithm predicted that the difference of the gage outputs should be modeled using the intercept and the normal force. The sum of the two gage outputs, on the other hand, should be modeled using the intercept, the pitching moment, and the square of the pitching moment. Equations of the deflection of a cantilever beam are used to show that the search algorithm s two recommended math models can also be obtained after performing a rigorous theoretical analysis of the deflection of the sting balance under load. The analysis of the sting balance calibration data set is a rare example of a situation when regression models of balance calibration data can directly be derived from first principles of physics and engineering. In addition, it is interesting to see that the search algorithm recommended the same regression models for the data analysis using only a set of statistical quality metrics.

Modeling Group Differences in OLS and Orthogonal Regression: Implications for Differential Validity Studies

ERIC Educational Resources Information Center

Kane, Michael T.; Mroch, Andrew A.

2010-01-01

In evaluating the relationship between two measures across different groups (i.e., in evaluating "differential validity") it is necessary to examine differences in correlation coefficients and in regression lines. Ordinary least squares (OLS) regression is the standard method for fitting lines to data, but its criterion for optimal fit…

Tutorial on Using Regression Models with Count Outcomes Using R

ERIC Educational Resources Information Center

Beaujean, A. Alexander; Morgan, Grant B.

2016-01-01

Education researchers often study count variables, such as times a student reached a goal, discipline referrals, and absences. Most researchers that study these variables use typical regression methods (i.e., ordinary least-squares) either with or without transforming the count variables. In either case, using typical regression for count data can…

The crux of the method: assumptions in ordinary least squares and logistic regression.

PubMed

Long, Rebecca G

2008-10-01

Logistic regression has increasingly become the tool of choice when analyzing data with a binary dependent variable. While resources relating to the technique are widely available, clear discussions of why logistic regression should be used in place of ordinary least squares regression are difficult to find. The current paper compares and contrasts the assumptions of ordinary least squares with those of logistic regression and explains why logistic regression's looser assumptions make it adept at handling violations of the more important assumptions in ordinary least squares.

Hyperspectral Imaging for Predicting the Internal Quality of Kiwifruits Based on Variable Selection Algorithms and Chemometric Models.

PubMed

Zhu, Hongyan; Chu, Bingquan; Fan, Yangyang; Tao, Xiaoya; Yin, Wenxin; He, Yong

2017-08-10

We investigated the feasibility and potentiality of determining firmness, soluble solids content (SSC), and pH in kiwifruits using hyperspectral imaging, combined with variable selection methods and calibration models. The images were acquired by a push-broom hyperspectral reflectance imaging system covering two spectral ranges. Weighted regression coefficients (BW), successive projections algorithm (SPA) and genetic algorithm-partial least square (GAPLS) were compared and evaluated for the selection of effective wavelengths. Moreover, multiple linear regression (MLR), partial least squares regression and least squares support vector machine (LS-SVM) were developed to predict quality attributes quantitatively using effective wavelengths. The established models, particularly SPA-MLR, SPA-LS-SVM and GAPLS-LS-SVM, performed well. The SPA-MLR models for firmness (R pre  = 0.9812, RPD = 5.17) and SSC (R pre  = 0.9523, RPD = 3.26) at 380-1023 nm showed excellent performance, whereas GAPLS-LS-SVM was the optimal model at 874-1734 nm for predicting pH (R pre  = 0.9070, RPD = 2.60). Image processing algorithms were developed to transfer the predictive model in every pixel to generate prediction maps that visualize the spatial distribution of firmness and SSC. Hence, the results clearly demonstrated that hyperspectral imaging has the potential as a fast and non-invasive method to predict the quality attributes of kiwifruits.

Detection of Butter Adulteration with Lard by Employing (1)H-NMR Spectroscopy and Multivariate Data Analysis.

PubMed

Fadzillah, Nurrulhidayah Ahmad; Man, Yaakob bin Che; Rohman, Abdul; Rosman, Arieff Salleh; Ismail, Amin; Mustafa, Shuhaimi; Khatib, Alfi

2015-01-01

The authentication of food products from the presence of non-allowed components for certain religion like lard is very important. In this study, we used proton Nuclear Magnetic Resonance ((1)H-NMR) spectroscopy for the analysis of butter adulterated with lard by simultaneously quantification of all proton bearing compounds, and consequently all relevant sample classes. Since the spectra obtained were too complex to be analyzed visually by the naked eyes, the classification of spectra was carried out.The multivariate calibration of partial least square (PLS) regression was used for modelling the relationship between actual value of lard and predicted value. The model yielded a highest regression coefficient (R(2)) of 0.998 and the lowest root mean square error calibration (RMSEC) of 0.0091% and root mean square error prediction (RMSEP) of 0.0090, respectively. Cross validation testing evaluates the predictive power of the model. PLS model was shown as good models as the intercept of R(2)Y and Q(2)Y were 0.0853 and -0.309, respectively.

A quasi-Monte-Carlo comparison of parametric and semiparametric regression methods for heavy-tailed and non-normal data: an application to healthcare costs.

PubMed

Jones, Andrew M; Lomas, James; Moore, Peter T; Rice, Nigel

2016-10-01

We conduct a quasi-Monte-Carlo comparison of the recent developments in parametric and semiparametric regression methods for healthcare costs, both against each other and against standard practice. The population of English National Health Service hospital in-patient episodes for the financial year 2007-2008 (summed for each patient) is randomly divided into two equally sized subpopulations to form an estimation set and a validation set. Evaluating out-of-sample using the validation set, a conditional density approximation estimator shows considerable promise in forecasting conditional means, performing best for accuracy of forecasting and among the best four for bias and goodness of fit. The best performing model for bias is linear regression with square-root-transformed dependent variables, whereas a generalized linear model with square-root link function and Poisson distribution performs best in terms of goodness of fit. Commonly used models utilizing a log-link are shown to perform badly relative to other models considered in our comparison.

Testing the Hypothesis of a Homoscedastic Error Term in Simple, Nonparametric Regression

ERIC Educational Resources Information Center

Wilcox, Rand R.

2006-01-01

Consider the nonparametric regression model Y = m(X)+ [tau](X)[epsilon], where X and [epsilon] are independent random variables, [epsilon] has a median of zero and variance [sigma][squared], [tau] is some unknown function used to model heteroscedasticity, and m(X) is an unknown function reflecting some conditional measure of location associated…

Criterion for evaluating the predictive ability of nonlinear regression models without cross-validation.

PubMed

Kaneko, Hiromasa; Funatsu, Kimito

2013-09-23

We propose predictive performance criteria for nonlinear regression models without cross-validation. The proposed criteria are the determination coefficient and the root-mean-square error for the midpoints between k-nearest-neighbor data points. These criteria can be used to evaluate predictive ability after the regression models are updated, whereas cross-validation cannot be performed in such a situation. The proposed method is effective and helpful in handling big data when cross-validation cannot be applied. By analyzing data from numerical simulations and quantitative structural relationships, we confirm that the proposed criteria enable the predictive ability of the nonlinear regression models to be appropriately quantified.

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.

A New Test of Linear Hypotheses in OLS Regression under Heteroscedasticity of Unknown Form

ERIC Educational Resources Information Center

Cai, Li; Hayes, Andrew F.

2008-01-01

When the errors in an ordinary least squares (OLS) regression model are heteroscedastic, hypothesis tests involving the regression coefficients can have Type I error rates that are far from the nominal significance level. Asymptotically, this problem can be rectified with the use of a heteroscedasticity-consistent covariance matrix (HCCM)…

Landscape risk factors for Lyme disease in the eastern broadleaf forest province of the Hudson River valley and the effect of explanatory data classification resolution.

PubMed

Messier, Kyle P; Jackson, Laura E; White, Jennifer L; Hilborn, Elizabeth D

2015-01-01

This study assessed how landcover classification affects associations between landscape characteristics and Lyme disease rate. Landscape variables were derived from the National Land Cover Database (NLCD), including native classes (e.g., deciduous forest, developed low intensity) and aggregate classes (e.g., forest, developed). Percent of each landcover type, median income, and centroid coordinates were calculated by census tract. Regression results from individual and aggregate variable models were compared with the dispersion parameter-based R(2) (Rα(2)) and AIC. The maximum Rα(2) was 0.82 and 0.83 for the best aggregate and individual model, respectively. The AICs for the best models differed by less than 0.5%. The aggregate model variables included forest, developed, agriculture, agriculture-squared, y-coordinate, y-coordinate-squared, income and income-squared. The individual model variables included deciduous forest, deciduous forest-squared, developed low intensity, pasture, y-coordinate, y-coordinate-squared, income, and income-squared. Results indicate that regional landscape models for Lyme disease rate are robust to NLCD landcover classification resolution. Published by Elsevier Ltd.

Sparse partial least squares regression for simultaneous dimension reduction and variable selection

PubMed Central

Chun, Hyonho; Keleş, Sündüz

2010-01-01

Partial least squares regression has been an alternative to ordinary least squares for handling multicollinearity in several areas of scientific research since the 1960s. It has recently gained much attention in the analysis of high dimensional genomic data. We show that known asymptotic consistency of the partial least squares estimator for a univariate response does not hold with the very large p and small n paradigm. We derive a similar result for a multivariate response regression with partial least squares. We then propose a sparse partial least squares formulation which aims simultaneously to achieve good predictive performance and variable selection by producing sparse linear combinations of the original predictors. We provide an efficient implementation of sparse partial least squares regression and compare it with well-known variable selection and dimension reduction approaches via simulation experiments. We illustrate the practical utility of sparse partial least squares regression in a joint analysis of gene expression and genomewide binding data. PMID:20107611

Inter-class sparsity based discriminative least square regression.

PubMed

Wen, Jie; Xu, Yong; Li, Zuoyong; Ma, Zhongli; Xu, Yuanrong

2018-06-01

Least square regression is a very popular supervised classification method. However, two main issues greatly limit its performance. The first one is that it only focuses on fitting the input features to the corresponding output labels while ignoring the correlations among samples. The second one is that the used label matrix, i.e., zero-one label matrix is inappropriate for classification. To solve these problems and improve the performance, this paper presents a novel method, i.e., inter-class sparsity based discriminative least square regression (ICS_DLSR), for multi-class classification. Different from other methods, the proposed method pursues that the transformed samples have a common sparsity structure in each class. For this goal, an inter-class sparsity constraint is introduced to the least square regression model such that the margins of samples from the same class can be greatly reduced while those of samples from different classes can be enlarged. In addition, an error term with row-sparsity constraint is introduced to relax the strict zero-one label matrix, which allows the method to be more flexible in learning the discriminative transformation matrix. These factors encourage the method to learn a more compact and discriminative transformation for regression and thus has the potential to perform better than other methods. Extensive experimental results show that the proposed method achieves the best performance in comparison with other methods for multi-class classification. Copyright © 2018 Elsevier Ltd. All rights reserved.

Exact and Approximate Statistical Inference for Nonlinear Regression and the Estimating Equation Approach.

PubMed

Demidenko, Eugene

2017-09-01

The exact density distribution of the nonlinear least squares estimator in the one-parameter regression model is derived in closed form and expressed through the cumulative distribution function of the standard normal variable. Several proposals to generalize this result are discussed. The exact density is extended to the estimating equation (EE) approach and the nonlinear regression with an arbitrary number of linear parameters and one intrinsically nonlinear parameter. For a very special nonlinear regression model, the derived density coincides with the distribution of the ratio of two normally distributed random variables previously obtained by Fieller (1932), unlike other approximations previously suggested by other authors. Approximations to the density of the EE estimators are discussed in the multivariate case. Numerical complications associated with the nonlinear least squares are illustrated, such as nonexistence and/or multiple solutions, as major factors contributing to poor density approximation. The nonlinear Markov-Gauss theorem is formulated based on the near exact EE density approximation.

ToxiM: A Toxicity Prediction Tool for Small Molecules Developed Using Machine Learning and Chemoinformatics Approaches.

PubMed

Sharma, Ashok K; Srivastava, Gopal N; Roy, Ankita; Sharma, Vineet K

2017-01-01

The experimental methods for the prediction of molecular toxicity are tedious and time-consuming tasks. Thus, the computational approaches could be used to develop alternative methods for toxicity prediction. We have developed a tool for the prediction of molecular toxicity along with the aqueous solubility and permeability of any molecule/metabolite. Using a comprehensive and curated set of toxin molecules as a training set, the different chemical and structural based features such as descriptors and fingerprints were exploited for feature selection, optimization and development of machine learning based classification and regression models. The compositional differences in the distribution of atoms were apparent between toxins and non-toxins, and hence, the molecular features were used for the classification and regression. On 10-fold cross-validation, the descriptor-based, fingerprint-based and hybrid-based classification models showed similar accuracy (93%) and Matthews's correlation coefficient (0.84). The performances of all the three models were comparable (Matthews's correlation coefficient = 0.84-0.87) on the blind dataset. In addition, the regression-based models using descriptors as input features were also compared and evaluated on the blind dataset. Random forest based regression model for the prediction of solubility performed better ( R 2 = 0.84) than the multi-linear regression (MLR) and partial least square regression (PLSR) models, whereas, the partial least squares based regression model for the prediction of permeability (caco-2) performed better ( R 2 = 0.68) in comparison to the random forest and MLR based regression models. The performance of final classification and regression models was evaluated using the two validation datasets including the known toxins and commonly used constituents of health products, which attests to its accuracy. The ToxiM web server would be a highly useful and reliable tool for the prediction of toxicity, solubility, and permeability of small molecules.

ToxiM: A Toxicity Prediction Tool for Small Molecules Developed Using Machine Learning and Chemoinformatics Approaches

PubMed Central

Sharma, Ashok K.; Srivastava, Gopal N.; Roy, Ankita; Sharma, Vineet K.

2017-01-01

The experimental methods for the prediction of molecular toxicity are tedious and time-consuming tasks. Thus, the computational approaches could be used to develop alternative methods for toxicity prediction. We have developed a tool for the prediction of molecular toxicity along with the aqueous solubility and permeability of any molecule/metabolite. Using a comprehensive and curated set of toxin molecules as a training set, the different chemical and structural based features such as descriptors and fingerprints were exploited for feature selection, optimization and development of machine learning based classification and regression models. The compositional differences in the distribution of atoms were apparent between toxins and non-toxins, and hence, the molecular features were used for the classification and regression. On 10-fold cross-validation, the descriptor-based, fingerprint-based and hybrid-based classification models showed similar accuracy (93%) and Matthews's correlation coefficient (0.84). The performances of all the three models were comparable (Matthews's correlation coefficient = 0.84–0.87) on the blind dataset. In addition, the regression-based models using descriptors as input features were also compared and evaluated on the blind dataset. Random forest based regression model for the prediction of solubility performed better (R2 = 0.84) than the multi-linear regression (MLR) and partial least square regression (PLSR) models, whereas, the partial least squares based regression model for the prediction of permeability (caco-2) performed better (R2 = 0.68) in comparison to the random forest and MLR based regression models. The performance of final classification and regression models was evaluated using the two validation datasets including the known toxins and commonly used constituents of health products, which attests to its accuracy. The ToxiM web server would be a highly useful and reliable tool for the prediction of toxicity, solubility, and permeability of small molecules. PMID:29249969

Modeling the language learning strategies and English language proficiency of pre-university students in UMS: A case study

NASA Astrophysics Data System (ADS)

Kiram, J. J.; Sulaiman, J.; Swanto, S.; Din, W. A.

2015-10-01

This study aims to construct a mathematical model of the relationship between a student's Language Learning Strategy usage and English Language proficiency. Fifty-six pre-university students of University Malaysia Sabah participated in this study. A self-report questionnaire called the Strategy Inventory for Language Learning was administered to them to measure their language learning strategy preferences before they sat for the Malaysian University English Test (MUET), the results of which were utilised to measure their English language proficiency. We attempted the model assessment specific to Multiple Linear Regression Analysis subject to variable selection using Stepwise regression. We conducted various assessments to the model obtained, including the Global F-test, Root Mean Square Error and R-squared. The model obtained suggests that not all language learning strategies should be included in the model in an attempt to predict Language Proficiency.

Buying a Better Air Force

DTIC Science & Technology

2006-03-01

identify if an explanatory variable may have been omitted due to model misspecification ( Ramsey , 1979). The RESET test resulted in failure to...Prob > F 0.0094 This model was also regressed using Huber-White estimators. Again, the Ramsey RESET test was done to ensure relevant...Aircraft. Annapolis, MD: Naval Institute Press, 2004. Ramsey , J. B. “ Tests for Specification Errors in Classical Least-Squares Regression Analysis

Evaluation of Ordinary Least Square (OLS) and Geographically Weighted Regression (GWR) for Water Quality Monitoring: A Case Study for the Estimation of Salinity

NASA Astrophysics Data System (ADS)

Nazeer, Majid; Bilal, Muhammad

2018-04-01

Landsat-5 Thematic Mapper (TM) dataset have been used to estimate salinity in the coastal area of Hong Kong. Four adjacent Landsat TM images were used in this study, which was atmospherically corrected using the Second Simulation of the Satellite Signal in the Solar Spectrum (6S) radiative transfer code. The atmospherically corrected images were further used to develop models for salinity using Ordinary Least Square (OLS) regression and Geographically Weighted Regression (GWR) based on in situ data of October 2009. Results show that the coefficient of determination ( R 2) of 0.42 between the OLS estimated and in situ measured salinity is much lower than that of the GWR model, which is two times higher ( R 2 = 0.86). It indicates that the GWR model has more ability than the OLS regression model to predict salinity and show its spatial heterogeneity better. It was observed that the salinity was high in Deep Bay (north-western part of Hong Kong) which might be due to the industrial waste disposal, whereas the salinity was estimated to be constant (32 practical salinity units) towards the open sea.

Spectral distance decay: Assessing species beta-diversity by quantile regression

USGS Publications Warehouse

Rocchinl, D.; Nagendra, H.; Ghate, R.; Cade, B.S.

2009-01-01

Remotely sensed data represents key information for characterizing and estimating biodiversity. Spectral distance among sites has proven to be a powerful approach for detecting species composition variability. Regression analysis of species similarity versus spectral distance may allow us to quantitatively estimate how beta-diversity in species changes with respect to spectral and ecological variability. In classical regression analysis, the residual sum of squares is minimized for the mean of the dependent variable distribution. However, many ecological datasets are characterized by a high number of zeroes that can add noise to the regression model. Quantile regression can be used to evaluate trend in the upper quantiles rather than a mean trend across the whole distribution of the dependent variable. In this paper, we used ordinary least square (ols) and quantile regression to estimate the decay of species similarity versus spectral distance. The achieved decay rates were statistically nonzero (p < 0.05) considering both ols and quantile regression. Nonetheless, ols regression estimate of mean decay rate was only half the decay rate indicated by the upper quantiles. Moreover, the intercept value, representing the similarity reached when spectral distance approaches zero, was very low compared with the intercepts of upper quantiles, which detected high species similarity when habitats are more similar. In this paper we demonstrated the power of using quantile regressions applied to spectral distance decay in order to reveal species diversity patterns otherwise lost or underestimated by ordinary least square regression. ?? 2009 American Society for Photogrammetry and Remote Sensing.

Production of deerbrush and mountain whitethorn related to shrub volume and overstory crown closure

Treesearch

John G. Kie

1985-01-01

Annual production by deerbrush (Ceanothus integerrimus) and mountain whitethorn shrubs (C. cordulatus) in the south-central Sierra Nevada of California was related to shrub volume, volume squared, and overstory crown closure by regression models. production increased as shrub volume and volume squared increased, and decreased as...

An iteratively reweighted least-squares approach to adaptive robust adjustment of parameters in linear regression models with autoregressive and t-distributed deviations

NASA Astrophysics Data System (ADS)

Kargoll, Boris; Omidalizarandi, Mohammad; Loth, Ina; Paffenholz, Jens-André; Alkhatib, Hamza

2018-03-01

In this paper, we investigate a linear regression time series model of possibly outlier-afflicted observations and autocorrelated random deviations. This colored noise is represented by a covariance-stationary autoregressive (AR) process, in which the independent error components follow a scaled (Student's) t-distribution. This error model allows for the stochastic modeling of multiple outliers and for an adaptive robust maximum likelihood (ML) estimation of the unknown regression and AR coefficients, the scale parameter, and the degree of freedom of the t-distribution. This approach is meant to be an extension of known estimators, which tend to focus only on the regression model, or on the AR error model, or on normally distributed errors. For the purpose of ML estimation, we derive an expectation conditional maximization either algorithm, which leads to an easy-to-implement version of iteratively reweighted least squares. The estimation performance of the algorithm is evaluated via Monte Carlo simulations for a Fourier as well as a spline model in connection with AR colored noise models of different orders and with three different sampling distributions generating the white noise components. We apply the algorithm to a vibration dataset recorded by a high-accuracy, single-axis accelerometer, focusing on the evaluation of the estimated AR colored noise model.

Hyperspectral analysis of soil organic matter in coal mining regions using wavelets, correlations, and partial least squares regression.

PubMed

Lin, Lixin; Wang, Yunjia; Teng, Jiyao; Wang, Xuchen

2016-02-01

Hyperspectral estimation of soil organic matter (SOM) in coal mining regions is an important tool for enhancing fertilization in soil restoration programs. The correlation--partial least squares regression (PLSR) method effectively solves the information loss problem of correlation--multiple linear stepwise regression, but results of the correlation analysis must be optimized to improve precision. This study considers the relationship between spectral reflectance and SOM based on spectral reflectance curves of soil samples collected from coal mining regions. Based on the major absorption troughs in the 400-1006 nm spectral range, PLSR analysis was performed using 289 independent bands of the second derivative (SDR) with three levels and measured SOM values. A wavelet-correlation-PLSR (W-C-PLSR) model was then constructed. By amplifying useful information that was previously obscured by noise, the W-C-PLSR model was optimal for estimating SOM content, with smaller prediction errors in both calibration (R(2) = 0.970, root mean square error (RMSEC) = 3.10, and mean relative error (MREC) = 8.75) and validation (RMSEV = 5.85 and MREV = 14.32) analyses, as compared with other models. Results indicate that W-C-PLSR has great potential to estimate SOM in coal mining regions.

Spectroscopic Determination of Aboveground Biomass in Grasslands Using Spectral Transformations, Support Vector Machine and Partial Least Squares Regression

PubMed Central

Marabel, Miguel; Alvarez-Taboada, Flor

2013-01-01

Aboveground biomass (AGB) is one of the strategic biophysical variables of interest in vegetation studies. The main objective of this study was to evaluate the Support Vector Machine (SVM) and Partial Least Squares Regression (PLSR) for estimating the AGB of grasslands from field spectrometer data and to find out which data pre-processing approach was the most suitable. The most accurate model to predict the total AGB involved PLSR and the Maximum Band Depth index derived from the continuum removed reflectance in the absorption features between 916–1,120 nm and 1,079–1,297 nm (R2 = 0.939, RMSE = 7.120 g/m2). Regarding the green fraction of the AGB, the Area Over the Minimum index derived from the continuum removed spectra provided the most accurate model overall (R2 = 0.939, RMSE = 3.172 g/m2). Identifying the appropriate absorption features was proved to be crucial to improve the performance of PLSR to estimate the total and green aboveground biomass, by using the indices derived from those spectral regions. Ordinary Least Square Regression could be used as a surrogate for the PLSR approach with the Area Over the Minimum index as the independent variable, although the resulting model would not be as accurate. PMID:23925082

Near-infrared hyperspectral imaging and partial least squares regression for rapid and reagentless determination of Enterobacteriaceae on chicken fillets.

PubMed

Feng, Yao-Ze; Elmasry, Gamal; Sun, Da-Wen; Scannell, Amalia G M; Walsh, Des; Morcy, Noha

2013-06-01

Bacterial pathogens are the main culprits for outbreaks of food-borne illnesses. This study aimed to use the hyperspectral imaging technique as a non-destructive tool for quantitative and direct determination of Enterobacteriaceae loads on chicken fillets. Partial least squares regression (PLSR) models were established and the best model using full wavelengths was obtained in the spectral range 930-1450 nm with coefficients of determination R(2)≥ 0.82 and root mean squared errors (RMSEs) ≤ 0.47 log(10)CFUg(-1). In further development of simplified models, second derivative spectra and weighted PLS regression coefficients (BW) were utilised to select important wavelengths. However, the three wavelengths (930, 1121 and 1345 nm) selected from BW were competent and more preferred for predicting Enterobacteriaceae loads with R(2) of 0.89, 0.86 and 0.87 and RMSEs of 0.33, 0.40 and 0.45 log(10)CFUg(-1) for calibration, cross-validation and prediction, respectively. Besides, the constructed prediction map provided the distribution of Enterobacteriaceae bacteria on chicken fillets, which cannot be achieved by conventional methods. It was demonstrated that hyperspectral imaging is a potential tool for determining food sanitation and detecting bacterial pathogens on food matrix without using complicated laboratory regimes. Copyright © 2012 Elsevier Ltd. All rights reserved.

Method for nonlinear exponential regression analysis

NASA Technical Reports Server (NTRS)

Junkin, B. G.

1972-01-01

Two computer programs developed according to two general types of exponential models for conducting nonlinear exponential regression analysis are described. Least squares procedure is used in which the nonlinear problem is linearized by expanding in a Taylor series. Program is written in FORTRAN 5 for the Univac 1108 computer.

Fast function-on-scalar regression with penalized basis expansions.

PubMed

Reiss, Philip T; Huang, Lei; Mennes, Maarten

2010-01-01

Regression models for functional responses and scalar predictors are often fitted by means of basis functions, with quadratic roughness penalties applied to avoid overfitting. The fitting approach described by Ramsay and Silverman in the 1990 s amounts to a penalized ordinary least squares (P-OLS) estimator of the coefficient functions. We recast this estimator as a generalized ridge regression estimator, and present a penalized generalized least squares (P-GLS) alternative. We describe algorithms by which both estimators can be implemented, with automatic selection of optimal smoothing parameters, in a more computationally efficient manner than has heretofore been available. We discuss pointwise confidence intervals for the coefficient functions, simultaneous inference by permutation tests, and model selection, including a novel notion of pointwise model selection. P-OLS and P-GLS are compared in a simulation study. Our methods are illustrated with an analysis of age effects in a functional magnetic resonance imaging data set, as well as a reanalysis of a now-classic Canadian weather data set. An R package implementing the methods is publicly available.

Quantification of brain lipids by FTIR spectroscopy and partial least squares regression

NASA Astrophysics Data System (ADS)

Dreissig, Isabell; Machill, Susanne; Salzer, Reiner; Krafft, Christoph

2009-01-01

Brain tissue is characterized by high lipid content. Its content decreases and the lipid composition changes during transformation from normal brain tissue to tumors. Therefore, the analysis of brain lipids might complement the existing diagnostic tools to determine the tumor type and tumor grade. Objective of this work is to extract lipids from gray matter and white matter of porcine brain tissue, record infrared (IR) spectra of these extracts and develop a quantification model for the main lipids based on partial least squares (PLS) regression. IR spectra of the pure lipids cholesterol, cholesterol ester, phosphatidic acid, phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, phosphatidylinositol, sphingomyelin, galactocerebroside and sulfatide were used as references. Two lipid mixtures were prepared for training and validation of the quantification model. The composition of lipid extracts that were predicted by the PLS regression of IR spectra was compared with lipid quantification by thin layer chromatography.

A comparison of different strategies in multivariate regression models for the direct determination of Mn, Cr, and Ni in steel samples using laser-induced breakdown spectroscopy

NASA Astrophysics Data System (ADS)

Luna, Aderval S.; Gonzaga, Fabiano B.; da Rocha, Werickson F. C.; Lima, Igor C. A.

2018-01-01

Laser-induced breakdown spectroscopy (LIBS) analysis was carried out on eleven steel samples to quantify the concentrations of chromium, nickel, and manganese. LIBS spectral data were correlated to known concentrations of the samples using different strategies in partial least squares (PLS) regression models. For the PLS analysis, one predictive model was separately generated for each element, while different approaches were used for the selection of variables (VIP: variable importance in projection and iPLS: interval partial least squares) in the PLS model to quantify the contents of the elements. The comparison of the performance of the models showed that there was no significant statistical difference using the Wilcoxon signed rank test. The elliptical joint confidence region (EJCR) did not detect systematic errors in these proposed methodologies for each metal.

Spatial Autocorrelation Approaches to Testing Residuals from Least Squares Regression.

PubMed

Chen, Yanguang

2016-01-01

In geo-statistics, the Durbin-Watson test is frequently employed to detect the presence of residual serial correlation from least squares regression analyses. However, the Durbin-Watson statistic is only suitable for ordered time or spatial series. If the variables comprise cross-sectional data coming from spatial random sampling, the test will be ineffectual because the value of Durbin-Watson's statistic depends on the sequence of data points. This paper develops two new statistics for testing serial correlation of residuals from least squares regression based on spatial samples. By analogy with the new form of Moran's index, an autocorrelation coefficient is defined with a standardized residual vector and a normalized spatial weight matrix. Then by analogy with the Durbin-Watson statistic, two types of new serial correlation indices are constructed. As a case study, the two newly presented statistics are applied to a spatial sample of 29 China's regions. These results show that the new spatial autocorrelation models can be used to test the serial correlation of residuals from regression analysis. In practice, the new statistics can make up for the deficiencies of the Durbin-Watson test.

Application of near-infrared spectroscopy for the rapid quality assessment of Radix Paeoniae Rubra

NASA Astrophysics Data System (ADS)

Zhan, Hao; Fang, Jing; Tang, Liying; Yang, Hongjun; Li, Hua; Wang, Zhuju; Yang, Bin; Wu, Hongwei; Fu, Meihong

2017-08-01

Near-infrared (NIR) spectroscopy with multivariate analysis was used to quantify gallic acid, catechin, albiflorin, and paeoniflorin in Radix Paeoniae Rubra, and the feasibility to classify the samples originating from different areas was investigated. A new high-performance liquid chromatography method was developed and validated to analyze gallic acid, catechin, albiflorin, and paeoniflorin in Radix Paeoniae Rubra as the reference. Partial least squares (PLS), principal component regression (PCR), and stepwise multivariate linear regression (SMLR) were performed to calibrate the regression model. Different data pretreatments such as derivatives (1st and 2nd), multiplicative scatter correction, standard normal variate, Savitzky-Golay filter, and Norris derivative filter were applied to remove the systematic errors. The performance of the model was evaluated according to the root mean square of calibration (RMSEC), root mean square error of prediction (RMSEP), root mean square error of cross-validation (RMSECV), and correlation coefficient (r). The results show that compared to PCR and SMLR, PLS had a lower RMSEC, RMSECV, and RMSEP and higher r for all the four analytes. PLS coupled with proper pretreatments showed good performance in both the fitting and predicting results. Furthermore, the original areas of Radix Paeoniae Rubra samples were partly distinguished by principal component analysis. This study shows that NIR with PLS is a reliable, inexpensive, and rapid tool for the quality assessment of Radix Paeoniae Rubra.

The comparison of robust partial least squares regression with robust principal component regression on a real

NASA Astrophysics Data System (ADS)

Polat, Esra; Gunay, Suleyman

2013-10-01

One of the problems encountered in Multiple Linear Regression (MLR) is multicollinearity, which causes the overestimation of the regression parameters and increase of the variance of these parameters. Hence, in case of multicollinearity presents, biased estimation procedures such as classical Principal Component Regression (CPCR) and Partial Least Squares Regression (PLSR) are then performed. SIMPLS algorithm is the leading PLSR algorithm because of its speed, efficiency and results are easier to interpret. However, both of the CPCR and SIMPLS yield very unreliable results when the data set contains outlying observations. Therefore, Hubert and Vanden Branden (2003) have been presented a robust PCR (RPCR) method and a robust PLSR (RPLSR) method called RSIMPLS. In RPCR, firstly, a robust Principal Component Analysis (PCA) method for high-dimensional data on the independent variables is applied, then, the dependent variables are regressed on the scores using a robust regression method. RSIMPLS has been constructed from a robust covariance matrix for high-dimensional data and robust linear regression. The purpose of this study is to show the usage of RPCR and RSIMPLS methods on an econometric data set, hence, making a comparison of two methods on an inflation model of Turkey. The considered methods have been compared in terms of predictive ability and goodness of fit by using a robust Root Mean Squared Error of Cross-validation (R-RMSECV), a robust R2 value and Robust Component Selection (RCS) statistic.

Developing a dengue forecast model using machine learning: A case study in China.

PubMed

Guo, Pi; Liu, Tao; Zhang, Qin; Wang, Li; Xiao, Jianpeng; Zhang, Qingying; Luo, Ganfeng; Li, Zhihao; He, Jianfeng; Zhang, Yonghui; Ma, Wenjun

2017-10-01

In China, dengue remains an important public health issue with expanded areas and increased incidence recently. Accurate and timely forecasts of dengue incidence in China are still lacking. We aimed to use the state-of-the-art machine learning algorithms to develop an accurate predictive model of dengue. Weekly dengue cases, Baidu search queries and climate factors (mean temperature, relative humidity and rainfall) during 2011-2014 in Guangdong were gathered. A dengue search index was constructed for developing the predictive models in combination with climate factors. The observed year and week were also included in the models to control for the long-term trend and seasonality. Several machine learning algorithms, including the support vector regression (SVR) algorithm, step-down linear regression model, gradient boosted regression tree algorithm (GBM), negative binomial regression model (NBM), least absolute shrinkage and selection operator (LASSO) linear regression model and generalized additive model (GAM), were used as candidate models to predict dengue incidence. Performance and goodness of fit of the models were assessed using the root-mean-square error (RMSE) and R-squared measures. The residuals of the models were examined using the autocorrelation and partial autocorrelation function analyses to check the validity of the models. The models were further validated using dengue surveillance data from five other provinces. The epidemics during the last 12 weeks and the peak of the 2014 large outbreak were accurately forecasted by the SVR model selected by a cross-validation technique. Moreover, the SVR model had the consistently smallest prediction error rates for tracking the dynamics of dengue and forecasting the outbreaks in other areas in China. The proposed SVR model achieved a superior performance in comparison with other forecasting techniques assessed in this study. The findings can help the government and community respond early to dengue epidemics.

Modelling daily dissolved oxygen concentration using least square support vector machine, multivariate adaptive regression splines and M5 model tree

NASA Astrophysics Data System (ADS)

Heddam, Salim; Kisi, Ozgur

2018-04-01

In the present study, three types of artificial intelligence techniques, least square support vector machine (LSSVM), multivariate adaptive regression splines (MARS) and M5 model tree (M5T) are applied for modeling daily dissolved oxygen (DO) concentration using several water quality variables as inputs. The DO concentration and water quality variables data from three stations operated by the United States Geological Survey (USGS) were used for developing the three models. The water quality data selected consisted of daily measured of water temperature (TE, °C), pH (std. unit), specific conductance (SC, μS/cm) and discharge (DI cfs), are used as inputs to the LSSVM, MARS and M5T models. The three models were applied for each station separately and compared to each other. According to the results obtained, it was found that: (i) the DO concentration could be successfully estimated using the three models and (ii) the best model among all others differs from one station to another.

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.

[Gaussian process regression and its application in near-infrared spectroscopy analysis].

PubMed

Feng, Ai-Ming; Fang, Li-Min; Lin, Min

2011-06-01

Gaussian process (GP) is applied in the present paper as a chemometric method to explore the complicated relationship between the near infrared (NIR) spectra and ingredients. After the outliers were detected by Monte Carlo cross validation (MCCV) method and removed from dataset, different preprocessing methods, such as multiplicative scatter correction (MSC), smoothing and derivate, were tried for the best performance of the models. Furthermore, uninformative variable elimination (UVE) was introduced as a variable selection technique and the characteristic wavelengths obtained were further employed as input for modeling. A public dataset with 80 NIR spectra of corn was introduced as an example for evaluating the new algorithm. The optimal models for oil, starch and protein were obtained by the GP regression method. The performance of the final models were evaluated according to the root mean square error of calibration (RMSEC), root mean square error of cross-validation (RMSECV), root mean square error of prediction (RMSEP) and correlation coefficient (r). The models give good calibration ability with r values above 0.99 and the prediction ability is also satisfactory with r values higher than 0.96. The overall results demonstrate that GP algorithm is an effective chemometric method and is promising for the NIR analysis.

Post-processing through linear regression

NASA Astrophysics Data System (ADS)

van Schaeybroeck, B.; Vannitsem, S.

2011-03-01

Various post-processing techniques are compared for both deterministic and ensemble forecasts, all based on linear regression between forecast data and observations. In order to evaluate the quality of the regression methods, three criteria are proposed, related to the effective correction of forecast error, the optimal variability of the corrected forecast and multicollinearity. The regression schemes under consideration include the ordinary least-square (OLS) method, a new time-dependent Tikhonov regularization (TDTR) method, the total least-square method, a new geometric-mean regression (GM), a recently introduced error-in-variables (EVMOS) method and, finally, a "best member" OLS method. The advantages and drawbacks of each method are clarified. These techniques are applied in the context of the 63 Lorenz system, whose model version is affected by both initial condition and model errors. For short forecast lead times, the number and choice of predictors plays an important role. Contrarily to the other techniques, GM degrades when the number of predictors increases. At intermediate lead times, linear regression is unable to provide corrections to the forecast and can sometimes degrade the performance (GM and the best member OLS with noise). At long lead times the regression schemes (EVMOS, TDTR) which yield the correct variability and the largest correlation between ensemble error and spread, should be preferred.

Improved estimation of PM2.5 using Lagrangian satellite-measured aerosol optical depth

NASA Astrophysics Data System (ADS)

Olivas Saunders, Rolando

Suspended particulate matter (aerosols) with aerodynamic diameters less than 2.5 mum (PM2.5) has negative effects on human health, plays an important role in climate change and also causes the corrosion of structures by acid deposition. Accurate estimates of PM2.5 concentrations are thus relevant in air quality, epidemiology, cloud microphysics and climate forcing studies. Aerosol optical depth (AOD) retrieved by the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite instrument has been used as an empirical predictor to estimate ground-level concentrations of PM2.5 . These estimates usually have large uncertainties and errors. The main objective of this work is to assess the value of using upwind (Lagrangian) MODIS-AOD as predictors in empirical models of PM2.5. The upwind locations of the Lagrangian AOD were estimated using modeled backward air trajectories. Since the specification of an arrival elevation is somewhat arbitrary, trajectories were calculated to arrive at four different elevations at ten measurement sites within the continental United States. A systematic examination revealed trajectory model calculations to be sensitive to starting elevation. With a 500 m difference in starting elevation, the 48-hr mean horizontal separation of trajectory endpoints was 326 km. When the difference in starting elevation was doubled and tripled to 1000 m and 1500m, the mean horizontal separation of trajectory endpoints approximately doubled and tripled to 627 km and 886 km, respectively. A seasonal dependence of this sensitivity was also found: the smallest mean horizontal separation of trajectory endpoints was exhibited during the summer and the largest separations during the winter. A daily average AOD product was generated and coupled to the trajectory model in order to determine AOD values upwind of the measurement sites during the period 2003-2007. Empirical models that included in situ AOD and upwind AOD as predictors of PM2.5 were generated by multivariate linear regressions using the least squares method. The multivariate models showed improved performance over the single variable regression (PM2.5 and in situ AOD) models. The statistical significance of the improvement of the multivariate models over the single variable regression models was tested using the extra sum of squares principle. In many cases, even when the R-squared was high for the multivariate models, the improvement over the single models was not statistically significant. The R-squared of these multivariate models varied with respect to seasons, with the best performance occurring during the summer months. A set of seasonal categorical variables was included in the regressions to exploit this variability. The multivariate regression models that included these categorical seasonal variables performed better than the models that didn't account for seasonal variability. Furthermore, 71% of these regressions exhibited improvement over the single variable models that was statistically significant at a 95% confidence level.

Accounting for measurement error in log regression models with applications to accelerated testing.

PubMed

Richardson, Robert; Tolley, H Dennis; Evenson, William E; Lunt, Barry M

2018-01-01

In regression settings, parameter estimates will be biased when the explanatory variables are measured with error. This bias can significantly affect modeling goals. In particular, accelerated lifetime testing involves an extrapolation of the fitted model, and a small amount of bias in parameter estimates may result in a significant increase in the bias of the extrapolated predictions. Additionally, bias may arise when the stochastic component of a log regression model is assumed to be multiplicative when the actual underlying stochastic component is additive. To account for these possible sources of bias, a log regression model with measurement error and additive error is approximated by a weighted regression model which can be estimated using Iteratively Re-weighted Least Squares. Using the reduced Eyring equation in an accelerated testing setting, the model is compared to previously accepted approaches to modeling accelerated testing data with both simulations and real data.

Multi-parameters monitoring during traditional Chinese medicine concentration process with near infrared spectroscopy and chemometrics

NASA Astrophysics Data System (ADS)

Liu, Ronghua; Sun, Qiaofeng; Hu, Tian; Li, Lian; Nie, Lei; Wang, Jiayue; Zhou, Wanhui; Zang, Hengchang

2018-03-01

As a powerful process analytical technology (PAT) tool, near infrared (NIR) spectroscopy has been widely used in real-time monitoring. In this study, NIR spectroscopy was applied to monitor multi-parameters of traditional Chinese medicine (TCM) Shenzhiling oral liquid during the concentration process to guarantee the quality of products. Five lab scale batches were employed to construct quantitative models to determine five chemical ingredients and physical change (samples density) during concentration process. The paeoniflorin, albiflorin, liquiritin and samples density were modeled by partial least square regression (PLSR), while the content of the glycyrrhizic acid and cinnamic acid were modeled by support vector machine regression (SVMR). Standard normal variate (SNV) and/or Savitzkye-Golay (SG) smoothing with derivative methods were adopted for spectra pretreatment. Variable selection methods including correlation coefficient (CC), competitive adaptive reweighted sampling (CARS) and interval partial least squares regression (iPLS) were performed for optimizing the models. The results indicated that NIR spectroscopy was an effective tool to successfully monitoring the concentration process of Shenzhiling oral liquid.

Soil sail content estimation in the yellow river delta with satellite hyperspectral data

USGS Publications Warehouse

Weng, Yongling; Gong, Peng; Zhu, Zhi-Liang

2008-01-01

Soil salinization is one of the most common land degradation processes and is a severe environmental hazard. The primary objective of this study is to investigate the potential of predicting salt content in soils with hyperspectral data acquired with EO-1 Hyperion. Both partial least-squares regression (PLSR) and conventional multiple linear regression (MLR), such as stepwise regression (SWR), were tested as the prediction model. PLSR is commonly used to overcome the problem caused by high-dimensional and correlated predictors. Chemical analysis of 95 samples collected from the top layer of soils in the Yellow River delta area shows that salt content was high on average, and the dominant chemicals in the saline soil were NaCl and MgCl2. Multivariate models were established between soil contents and hyperspectral data. Our results indicate that the PLSR technique with laboratory spectral data has a strong prediction capacity. Spectral bands at 1487-1527, 1971-1991, 2032-2092, and 2163-2355 nm possessed large absolute values of regression coefficients, with the largest coefficient at 2203 nm. We obtained a root mean squared error (RMSE) for calibration (with 61 samples) of RMSEC = 0.753 (R2 = 0.893) and a root mean squared error for validation (with 30 samples) of RMSEV = 0.574. The prediction model was applied on a pixel-by-pixel basis to a Hyperion reflectance image to yield a quantitative surface distribution map of soil salt content. The result was validated successfully from 38 sampling points. We obtained an RMSE estimate of 1.037 (R2 = 0.784) for the soil salt content map derived by the PLSR model. The salinity map derived from the SWR model shows that the predicted value is higher than the true value. These results demonstrate that the PLSR method is a more suitable technique than stepwise regression for quantitative estimation of soil salt content in a large area. ?? 2008 CASI.

Using ridge regression in systematic pointing error corrections

NASA Technical Reports Server (NTRS)

Guiar, C. N.

1988-01-01

A pointing error model is used in the antenna calibration process. Data from spacecraft or radio star observations are used to determine the parameters in the model. However, the regression variables are not truly independent, displaying a condition known as multicollinearity. Ridge regression, a biased estimation technique, is used to combat the multicollinearity problem. Two data sets pertaining to Voyager 1 spacecraft tracking (days 105 and 106 of 1987) were analyzed using both linear least squares and ridge regression methods. The advantages and limitations of employing the technique are presented. The problem is not yet fully resolved.

Quantum State Tomography via Linear Regression Estimation

PubMed Central

Qi, Bo; Hou, Zhibo; Li, Li; Dong, Daoyi; Xiang, Guoyong; Guo, Guangcan

2013-01-01

A simple yet efficient state reconstruction algorithm of linear regression estimation (LRE) is presented for quantum state tomography. In this method, quantum state reconstruction is converted into a parameter estimation problem of a linear regression model and the least-squares method is employed to estimate the unknown parameters. An asymptotic mean squared error (MSE) upper bound for all possible states to be estimated is given analytically, which depends explicitly upon the involved measurement bases. This analytical MSE upper bound can guide one to choose optimal measurement sets. The computational complexity of LRE is O(d4) where d is the dimension of the quantum state. Numerical examples show that LRE is much faster than maximum-likelihood estimation for quantum state tomography. PMID:24336519

Pan evaporation modeling using six different heuristic computing methods in different climates of China

NASA Astrophysics Data System (ADS)

Wang, Lunche; Kisi, Ozgur; Zounemat-Kermani, Mohammad; Li, Hui

2017-01-01

Pan evaporation (Ep) plays important roles in agricultural water resources management. One of the basic challenges is modeling Ep using limited climatic parameters because there are a number of factors affecting the evaporation rate. This study investigated the abilities of six different soft computing methods, multi-layer perceptron (MLP), generalized regression neural network (GRNN), fuzzy genetic (FG), least square support vector machine (LSSVM), multivariate adaptive regression spline (MARS), adaptive neuro-fuzzy inference systems with grid partition (ANFIS-GP), and two regression methods, multiple linear regression (MLR) and Stephens and Stewart model (SS) in predicting monthly Ep. Long-term climatic data at various sites crossing a wide range of climates during 1961-2000 are used for model development and validation. The results showed that the models have different accuracies in different climates and the MLP model performed superior to the other models in predicting monthly Ep at most stations using local input combinations (for example, the MAE (mean absolute errors), RMSE (root mean square errors), and determination coefficient (R2) are 0.314 mm/day, 0.405 mm/day and 0.988, respectively for HEB station), while GRNN model performed better in Tibetan Plateau (MAE, RMSE and R2 are 0.459 mm/day, 0.592 mm/day and 0.932, respectively). The accuracies of above models ranked as: MLP, GRNN, LSSVM, FG, ANFIS-GP, MARS and MLR. The overall results indicated that the soft computing techniques generally performed better than the regression methods, but MLR and SS models can be more preferred at some climatic zones instead of complex nonlinear models, for example, the BJ (Beijing), CQ (Chongqing) and HK (Haikou) stations. Therefore, it can be concluded that Ep could be successfully predicted using above models in hydrological modeling studies.

Robust Bayesian linear regression with application to an analysis of the CODATA values for the Planck constant

NASA Astrophysics Data System (ADS)

Wübbeler, Gerd; Bodnar, Olha; Elster, Clemens

2018-02-01

Weighted least-squares estimation is commonly applied in metrology to fit models to measurements that are accompanied with quoted uncertainties. The weights are chosen in dependence on the quoted uncertainties. However, when data and model are inconsistent in view of the quoted uncertainties, this procedure does not yield adequate results. When it can be assumed that all uncertainties ought to be rescaled by a common factor, weighted least-squares estimation may still be used, provided that a simple correction of the uncertainty obtained for the estimated model is applied. We show that these uncertainties and credible intervals are robust, as they do not rely on the assumption of a Gaussian distribution of the data. Hence, common software for weighted least-squares estimation may still safely be employed in such a case, followed by a simple modification of the uncertainties obtained by that software. We also provide means of checking the assumptions of such an approach. The Bayesian regression procedure is applied to analyze the CODATA values for the Planck constant published over the past decades in terms of three different models: a constant model, a straight line model and a spline model. Our results indicate that the CODATA values may not have yet stabilized.

Determinant of securitization asset pricing in Malaysia

NASA Astrophysics Data System (ADS)

Bakri, M. H.; Ali, R.; Ismail, S.; Sufian, F.; Baharom, A. H.

2014-12-01

Malaysian firms have been reported involve in Asset Back Securities since 1986s where Cagamas is a pioneer. This research aims to examine the factor influencing primary market spread. Least square method and regression analysis are applied for the study period 2004-2012. The result shows one determinants in internal regression model and three determinants in external regression influence or contribute to the primary market spread and are statistically significant in developing the securitization in Malaysia. It can be concluded that transaction size significantly contribute to the determinant primary market spread in internal regression model while liquidity, transaction size and crisis is significant in both regression model. From five hypotheses, three hypotheses support that the determinants have a relationship with primary market spread.

Acidity measurement of iron ore powders using laser-induced breakdown spectroscopy with partial least squares regression.

PubMed

Hao, Z Q; Li, C M; Shen, M; Yang, X Y; Li, K H; Guo, L B; Li, X Y; Lu, Y F; Zeng, X Y

2015-03-23

Laser-induced breakdown spectroscopy (LIBS) with partial least squares regression (PLSR) has been applied to measuring the acidity of iron ore, which can be defined by the concentrations of oxides: CaO, MgO, Al₂O₃, and SiO₂. With the conventional internal standard calibration, it is difficult to establish the calibration curves of CaO, MgO, Al₂O₃, and SiO₂ in iron ore due to the serious matrix effects. PLSR is effective to address this problem due to its excellent performance in compensating the matrix effects. In this work, fifty samples were used to construct the PLSR calibration models for the above-mentioned oxides. These calibration models were validated by the 10-fold cross-validation method with the minimum root-mean-square errors (RMSE). Another ten samples were used as a test set. The acidities were calculated according to the estimated concentrations of CaO, MgO, Al₂O₃, and SiO₂ using the PLSR models. The average relative error (ARE) and RMSE of the acidity achieved 3.65% and 0.0048, respectively, for the test samples.

Variable selection based on clustering analysis for improvement of polyphenols prediction in green tea using synchronous fluorescence spectra.

PubMed

Shan, Jiajia; Wang, Xue; Zhou, Hao; Han, Shuqing; Riza, Dimas Firmanda Al; Kondo, Naoshi

2018-03-13

Synchronous fluorescence spectra, combined with multivariate analysis were used to predict flavonoids content in green tea rapidly and nondestructively. This paper presented a new and efficient spectral intervals selection method called clustering based partial least square (CL-PLS), which selected informative wavelengths by combining clustering concept and partial least square (PLS) methods to improve models' performance by synchronous fluorescence spectra. The fluorescence spectra of tea samples were obtained and k-means and kohonen-self organizing map clustering algorithms were carried out to cluster full spectra into several clusters, and sub-PLS regression model was developed on each cluster. Finally, CL-PLS models consisting of gradually selected clusters were built. Correlation coefficient (R) was used to evaluate the effect on prediction performance of PLS models. In addition, variable influence on projection partial least square (VIP-PLS), selectivity ratio partial least square (SR-PLS), interval partial least square (iPLS) models and full spectra PLS model were investigated and the results were compared. The results showed that CL-PLS presented the best result for flavonoids prediction using synchronous fluorescence spectra.

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.

Noninvasive and fast measurement of blood glucose in vivo by near infrared (NIR) spectroscopy

NASA Astrophysics Data System (ADS)

Jintao, Xue; Liming, Ye; Yufei, Liu; Chunyan, Li; Han, Chen

2017-05-01

This research was to develop a method for noninvasive and fast blood glucose assay in vivo. Near-infrared (NIR) spectroscopy, a more promising technique compared to other methods, was investigated in rats with diabetes and normal rats. Calibration models are generated by two different multivariate strategies: partial least squares (PLS) as linear regression method and artificial neural networks (ANN) as non-linear regression method. The PLS model was optimized individually by considering spectral range, spectral pretreatment methods and number of model factors, while the ANN model was studied individually by selecting spectral pretreatment methods, parameters of network topology, number of hidden neurons, and times of epoch. The results of the validation showed the two models were robust, accurate and repeatable. Compared to the ANN model, the performance of the PLS model was much better, with lower root mean square error of validation (RMSEP) of 0.419 and higher correlation coefficients (R) of 96.22%.

Feasibility of using near infrared spectroscopy to detect and quantify an adulterant in high quality sandalwood oil.

PubMed

Kuriakose, Saji; Joe, I Hubert

2013-11-01

Determination of the authenticity of essential oils has become more significant, in recent years, following some illegal adulteration and contamination scandals. The present investigative study focuses on the application of near infrared spectroscopy to detect sample authenticity and quantify economic adulteration of sandalwood oils. Several data pre-treatments are investigated for calibration and prediction using partial least square regression (PLSR). The quantitative data analysis is done using a new spectral approach - full spectrum or sequential spectrum. The optimum number of PLS components is obtained according to the lowest root mean square error of calibration (RMSEC=0.00009% v/v). The lowest root mean square error of prediction (RMSEP=0.00016% v/v) in the test set and the highest coefficient of determination (R(2)=0.99989) are used as the evaluation tools for the best model. A nonlinear method, locally weighted regression (LWR), is added to extract nonlinear information and to compare with the linear PLSR model. Copyright © 2013 Elsevier B.V. All rights reserved.

Feasibility of using near infrared spectroscopy to detect and quantify an adulterant in high quality sandalwood oil

NASA Astrophysics Data System (ADS)

Kuriakose, Saji; Joe, I. Hubert

2013-11-01

Determination of the authenticity of essential oils has become more significant, in recent years, following some illegal adulteration and contamination scandals. The present investigative study focuses on the application of near infrared spectroscopy to detect sample authenticity and quantify economic adulteration of sandalwood oils. Several data pre-treatments are investigated for calibration and prediction using partial least square regression (PLSR). The quantitative data analysis is done using a new spectral approach - full spectrum or sequential spectrum. The optimum number of PLS components is obtained according to the lowest root mean square error of calibration (RMSEC = 0.00009% v/v). The lowest root mean square error of prediction (RMSEP = 0.00016% v/v) in the test set and the highest coefficient of determination (R2 = 0.99989) are used as the evaluation tools for the best model. A nonlinear method, locally weighted regression (LWR), is added to extract nonlinear information and to compare with the linear PLSR model.

Green method by diffuse reflectance infrared spectroscopy and spectral region selection for the quantification of sulphamethoxazole and trimethoprim in pharmaceutical formulations.

PubMed

da Silva, Fabiana E B; Flores, Érico M M; Parisotto, Graciele; Müller, Edson I; Ferrão, Marco F

2016-03-01

An alternative method for the quantification of sulphametoxazole (SMZ) and trimethoprim (TMP) using diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) and partial least square regression (PLS) was developed. Interval Partial Least Square (iPLS) and Synergy Partial Least Square (siPLS) were applied to select a spectral range that provided the lowest prediction error in comparison to the full-spectrum model. Fifteen commercial tablet formulations and forty-nine synthetic samples were used. The ranges of concentration considered were 400 to 900 mg g-1SMZ and 80 to 240 mg g-1 TMP. Spectral data were recorded between 600 and 4000 cm-1 with a 4 cm-1 resolution by Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS). The proposed procedure was compared to high performance liquid chromatography (HPLC). The results obtained from the root mean square error of prediction (RMSEP), during the validation of the models for samples of sulphamethoxazole (SMZ) and trimethoprim (TMP) using siPLS, demonstrate that this approach is a valid technique for use in quantitative analysis of pharmaceutical formulations. The selected interval algorithm allowed building regression models with minor errors when compared to the full spectrum PLS model. A RMSEP of 13.03 mg g-1for SMZ and 4.88 mg g-1 for TMP was obtained after the selection the best spectral regions by siPLS.

Improved accuracy in quantitative laser-induced breakdown spectroscopy using sub-models

USGS Publications Warehouse

Anderson, Ryan; Clegg, Samuel M.; Frydenvang, Jens; Wiens, Roger C.; McLennan, Scott M.; Morris, Richard V.; Ehlmann, Bethany L.; Dyar, M. Darby

2017-01-01

Accurate quantitative analysis of diverse geologic materials is one of the primary challenges faced by the Laser-Induced Breakdown Spectroscopy (LIBS)-based ChemCam instrument on the Mars Science Laboratory (MSL) rover. The SuperCam instrument on the Mars 2020 rover, as well as other LIBS instruments developed for geochemical analysis on Earth or other planets, will face the same challenge. Consequently, part of the ChemCam science team has focused on the development of improved multivariate analysis calibrations methods. Developing a single regression model capable of accurately determining the composition of very different target materials is difficult because the response of an element’s emission lines in LIBS spectra can vary with the concentration of other elements. We demonstrate a conceptually simple “sub-model” method for improving the accuracy of quantitative LIBS analysis of diverse target materials. The method is based on training several regression models on sets of targets with limited composition ranges and then “blending” these “sub-models” into a single final result. Tests of the sub-model method show improvement in test set root mean squared error of prediction (RMSEP) for almost all cases. The sub-model method, using partial least squares regression (PLS), is being used as part of the current ChemCam quantitative calibration, but the sub-model method is applicable to any multivariate regression method and may yield similar improvements.

Ordinary Least Squares and Quantile Regression: An Inquiry-Based Learning Approach to a Comparison of Regression Methods

ERIC Educational Resources Information Center

Helmreich, James E.; Krog, K. Peter

2018-01-01

We present a short, inquiry-based learning course on concepts and methods underlying ordinary least squares (OLS), least absolute deviation (LAD), and quantile regression (QR). Students investigate squared, absolute, and weighted absolute distance functions (metrics) as location measures. Using differential calculus and properties of convex…

Estimating riparian understory vegetation cover with beta regression and copula models

USGS Publications Warehouse

Eskelson, Bianca N.I.; Madsen, Lisa; Hagar, Joan C.; Temesgen, Hailemariam

2011-01-01

Understory vegetation communities are critical components of forest ecosystems. As a result, the importance of modeling understory vegetation characteristics in forested landscapes has become more apparent. Abundance measures such as shrub cover are bounded between 0 and 1, exhibit heteroscedastic error variance, and are often subject to spatial dependence. These distributional features tend to be ignored when shrub cover data are analyzed. The beta distribution has been used successfully to describe the frequency distribution of vegetation cover. Beta regression models ignoring spatial dependence (BR) and accounting for spatial dependence (BRdep) were used to estimate percent shrub cover as a function of topographic conditions and overstory vegetation structure in riparian zones in western Oregon. The BR models showed poor explanatory power (pseudo-R2 ≤ 0.34) but outperformed ordinary least-squares (OLS) and generalized least-squares (GLS) regression models with logit-transformed response in terms of mean square prediction error and absolute bias. We introduce a copula (COP) model that is based on the beta distribution and accounts for spatial dependence. A simulation study was designed to illustrate the effects of incorrectly assuming normality, equal variance, and spatial independence. It showed that BR, BRdep, and COP models provide unbiased parameter estimates, whereas OLS and GLS models result in slightly biased estimates for two of the three parameters. On the basis of the simulation study, 93–97% of the GLS, BRdep, and COP confidence intervals covered the true parameters, whereas OLS and BR only resulted in 84–88% coverage, which demonstrated the superiority of GLS, BRdep, and COP over OLS and BR models in providing standard errors for the parameter estimates in the presence of spatial dependence.

Comparing methods of analysing datasets with small clusters: case studies using four paediatric datasets.

PubMed

Marston, Louise; Peacock, Janet L; Yu, Keming; Brocklehurst, Peter; Calvert, Sandra A; Greenough, Anne; Marlow, Neil

2009-07-01

Studies of prematurely born infants contain a relatively large percentage of multiple births, so the resulting data have a hierarchical structure with small clusters of size 1, 2 or 3. Ignoring the clustering may lead to incorrect inferences. The aim of this study was to compare statistical methods which can be used to analyse such data: generalised estimating equations, multilevel models, multiple linear regression and logistic regression. Four datasets which differed in total size and in percentage of multiple births (n = 254, multiple 18%; n = 176, multiple 9%; n = 10 098, multiple 3%; n = 1585, multiple 8%) were analysed. With the continuous outcome, two-level models produced similar results in the larger dataset, while generalised least squares multilevel modelling (ML GLS 'xtreg' in Stata) and maximum likelihood multilevel modelling (ML MLE 'xtmixed' in Stata) produced divergent estimates using the smaller dataset. For the dichotomous outcome, most methods, except generalised least squares multilevel modelling (ML GH 'xtlogit' in Stata) gave similar odds ratios and 95% confidence intervals within datasets. For the continuous outcome, our results suggest using multilevel modelling. We conclude that generalised least squares multilevel modelling (ML GLS 'xtreg' in Stata) and maximum likelihood multilevel modelling (ML MLE 'xtmixed' in Stata) should be used with caution when the dataset is small. Where the outcome is dichotomous and there is a relatively large percentage of non-independent data, it is recommended that these are accounted for in analyses using logistic regression with adjusted standard errors or multilevel modelling. If, however, the dataset has a small percentage of clusters greater than size 1 (e.g. a population dataset of children where there are few multiples) there appears to be less need to adjust for clustering.

Advanced statistics: linear regression, part I: simple linear regression.

PubMed

Marill, Keith A

2004-01-01

Simple linear regression is a mathematical technique used to model the relationship between a single independent predictor variable and a single dependent outcome variable. In this, the first of a two-part series exploring concepts in linear regression analysis, the four fundamental assumptions and the mechanics of simple linear regression are reviewed. The most common technique used to derive the regression line, the method of least squares, is described. The reader will be acquainted with other important concepts in simple linear regression, including: variable transformations, dummy variables, relationship to inference testing, and leverage. Simplified clinical examples with small datasets and graphic models are used to illustrate the points. This will provide a foundation for the second article in this series: a discussion of multiple linear regression, in which there are multiple predictor variables.

On Using the Average Intercorrelation Among Predictor Variables and Eigenvector Orientation to Choose a Regression Solution.

ERIC Educational Resources Information Center

Mugrage, Beverly; And Others

Three ridge regression solutions are compared with ordinary least squares regression and with principal components regression using all components. Ridge regression, particularly the Lawless-Wang solution, out-performed ordinary least squares regression and the principal components solution on the criteria of stability of coefficient and closeness…

Demonstration of the Web-based Interspecies Correlation Estimation (Web-ICE) modeling application

EPA Science Inventory

The Web-based Interspecies Correlation Estimation (Web-ICE) modeling application is available to the risk assessment community through a user-friendly internet platform (http://epa.gov/ceampubl/fchain/webice/). ICE models are log-linear least square regressions that predict acute...

Using the Criterion-Predictor Factor Model to Compute the Probability of Detecting Prediction Bias with Ordinary Least Squares Regression

ERIC Educational Resources Information Center

Culpepper, Steven Andrew

2012-01-01

The study of prediction bias is important and the last five decades include research studies that examined whether test scores differentially predict academic or employment performance. Previous studies used ordinary least squares (OLS) to assess whether groups differ in intercepts and slopes. This study shows that OLS yields inaccurate inferences…

Technique for estimating the 2- to 500-year flood discharges on unregulated streams in rural Missouri

USGS Publications Warehouse

Alexander, Terry W.; Wilson, Gary L.

1995-01-01

A generalized least-squares regression technique was used to relate the 2- to 500-year flood discharges from 278 selected streamflow-gaging stations to statistically significant basin characteristics. The regression relations (estimating equations) were defined for three hydrologic regions (I, II, and III) in rural Missouri. Ordinary least-squares regression analyses indicate that drainage area (Regions I, II, and III) and main-channel slope (Regions I and II) are the only basin characteristics needed for computing the 2- to 500-year design-flood discharges at gaged or ungaged stream locations. The resulting generalized least-squares regression equations provide a technique for estimating the 2-, 5-, 10-, 25-, 50-, 100-, and 500-year flood discharges on unregulated streams in rural Missouri. The regression equations for Regions I and II were developed from stream-flow-gaging stations with drainage areas ranging from 0.13 to 11,500 square miles and 0.13 to 14,000 square miles, and main-channel slopes ranging from 1.35 to 150 feet per mile and 1.20 to 279 feet per mile. The regression equations for Region III were developed from streamflow-gaging stations with drainage areas ranging from 0.48 to 1,040 square miles. Standard errors of estimate for the generalized least-squares regression equations in Regions I, II, and m ranged from 30 to 49 percent.

Using Quantile and Asymmetric Least Squares Regression for Optimal Risk Adjustment.

PubMed

Lorenz, Normann

2017-06-01

In this paper, we analyze optimal risk adjustment for direct risk selection (DRS). Integrating insurers' activities for risk selection into a discrete choice model of individuals' health insurance choice shows that DRS has the structure of a contest. For the contest success function (csf) used in most of the contest literature (the Tullock-csf), optimal transfers for a risk adjustment scheme have to be determined by means of a restricted quantile regression, irrespective of whether insurers are primarily engaged in positive DRS (attracting low risks) or negative DRS (repelling high risks). This is at odds with the common practice of determining transfers by means of a least squares regression. However, this common practice can be rationalized for a new csf, but only if positive and negative DRSs are equally important; if they are not, optimal transfers have to be calculated by means of a restricted asymmetric least squares regression. Using data from German and Swiss health insurers, we find considerable differences between the three types of regressions. Optimal transfers therefore critically depend on which csf represents insurers' incentives for DRS and, if it is not the Tullock-csf, whether insurers are primarily engaged in positive or negative DRS. Copyright © 2016 John Wiley & Sons, Ltd. Copyright © 2016 John Wiley & Sons, Ltd.

Variable selection based on clustering analysis for improvement of polyphenols prediction in green tea using synchronous fluorescence spectra

NASA Astrophysics Data System (ADS)

Shan, Jiajia; Wang, Xue; Zhou, Hao; Han, Shuqing; Riza, Dimas Firmanda Al; Kondo, Naoshi

2018-04-01

Synchronous fluorescence spectra, combined with multivariate analysis were used to predict flavonoids content in green tea rapidly and nondestructively. This paper presented a new and efficient spectral intervals selection method called clustering based partial least square (CL-PLS), which selected informative wavelengths by combining clustering concept and partial least square (PLS) methods to improve models’ performance by synchronous fluorescence spectra. The fluorescence spectra of tea samples were obtained and k-means and kohonen-self organizing map clustering algorithms were carried out to cluster full spectra into several clusters, and sub-PLS regression model was developed on each cluster. Finally, CL-PLS models consisting of gradually selected clusters were built. Correlation coefficient (R) was used to evaluate the effect on prediction performance of PLS models. In addition, variable influence on projection partial least square (VIP-PLS), selectivity ratio partial least square (SR-PLS), interval partial least square (iPLS) models and full spectra PLS model were investigated and the results were compared. The results showed that CL-PLS presented the best result for flavonoids prediction using synchronous fluorescence spectra.

A Technique of Fuzzy C-Mean in Multiple Linear Regression Model toward Paddy Yield

NASA Astrophysics Data System (ADS)

Syazwan Wahab, Nur; Saifullah Rusiman, Mohd; Mohamad, Mahathir; Amira Azmi, Nur; Che Him, Norziha; Ghazali Kamardan, M.; Ali, Maselan

2018-04-01

In this paper, we propose a hybrid model which is a combination of multiple linear regression model and fuzzy c-means method. This research involved a relationship between 20 variates of the top soil that are analyzed prior to planting of paddy yields at standard fertilizer rates. Data used were from the multi-location trials for rice carried out by MARDI at major paddy granary in Peninsular Malaysia during the period from 2009 to 2012. Missing observations were estimated using mean estimation techniques. The data were analyzed using multiple linear regression model and a combination of multiple linear regression model and fuzzy c-means method. Analysis of normality and multicollinearity indicate that the data is normally scattered without multicollinearity among independent variables. Analysis of fuzzy c-means cluster the yield of paddy into two clusters before the multiple linear regression model can be used. The comparison between two method indicate that the hybrid of multiple linear regression model and fuzzy c-means method outperform the multiple linear regression model with lower value of mean square error.

A gentle introduction to quantile regression for ecologists

USGS Publications Warehouse

Cade, B.S.; Noon, B.R.

2003-01-01

Quantile regression is a way to estimate the conditional quantiles of a response variable distribution in the linear model that provides a more complete view of possible causal relationships between variables in ecological processes. Typically, all the factors that affect ecological processes are not measured and included in the statistical models used to investigate relationships between variables associated with those processes. As a consequence, there may be a weak or no predictive relationship between the mean of the response variable (y) distribution and the measured predictive factors (X). Yet there may be stronger, useful predictive relationships with other parts of the response variable distribution. This primer relates quantile regression estimates to prediction intervals in parametric error distribution regression models (eg least squares), and discusses the ordering characteristics, interval nature, sampling variation, weighting, and interpretation of the estimates for homogeneous and heterogeneous regression models.

Spatial Autocorrelation Approaches to Testing Residuals from Least Squares Regression

PubMed Central

Chen, Yanguang

2016-01-01

In geo-statistics, the Durbin-Watson test is frequently employed to detect the presence of residual serial correlation from least squares regression analyses. However, the Durbin-Watson statistic is only suitable for ordered time or spatial series. If the variables comprise cross-sectional data coming from spatial random sampling, the test will be ineffectual because the value of Durbin-Watson’s statistic depends on the sequence of data points. This paper develops two new statistics for testing serial correlation of residuals from least squares regression based on spatial samples. By analogy with the new form of Moran’s index, an autocorrelation coefficient is defined with a standardized residual vector and a normalized spatial weight matrix. Then by analogy with the Durbin-Watson statistic, two types of new serial correlation indices are constructed. As a case study, the two newly presented statistics are applied to a spatial sample of 29 China’s regions. These results show that the new spatial autocorrelation models can be used to test the serial correlation of residuals from regression analysis. In practice, the new statistics can make up for the deficiencies of the Durbin-Watson test. PMID:26800271

An interactive website for analytical method comparison and bias estimation.

PubMed

Bahar, Burak; Tuncel, Ayse F; Holmes, Earle W; Holmes, Daniel T

2017-12-01

Regulatory standards mandate laboratories to perform studies to ensure accuracy and reliability of their test results. Method comparison and bias estimation are important components of these studies. We developed an interactive website for evaluating the relative performance of two analytical methods using R programming language tools. The website can be accessed at https://bahar.shinyapps.io/method_compare/. The site has an easy-to-use interface that allows both copy-pasting and manual entry of data. It also allows selection of a regression model and creation of regression and difference plots. Available regression models include Ordinary Least Squares, Weighted-Ordinary Least Squares, Deming, Weighted-Deming, Passing-Bablok and Passing-Bablok for large datasets. The server processes the data and generates downloadable reports in PDF or HTML format. Our website provides clinical laboratories a practical way to assess the relative performance of two analytical methods. Copyright © 2017 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.

Improving Global Models of Remotely Sensed Ocean Chlorophyll Content Using Partial Least Squares and Geographically Weighted Regression

NASA Astrophysics Data System (ADS)

Gholizadeh, H.; Robeson, S. M.

2015-12-01

Empirical models have been widely used to estimate global chlorophyll content from remotely sensed data. Here, we focus on the standard NASA empirical models that use blue-green band ratios. These band ratio ocean color (OC) algorithms are in the form of fourth-order polynomials and the parameters of these polynomials (i.e. coefficients) are estimated from the NASA bio-Optical Marine Algorithm Data set (NOMAD). Most of the points in this data set have been sampled from tropical and temperate regions. However, polynomial coefficients obtained from this data set are used to estimate chlorophyll content in all ocean regions with different properties such as sea-surface temperature, salinity, and downwelling/upwelling patterns. Further, the polynomial terms in these models are highly correlated. In sum, the limitations of these empirical models are as follows: 1) the independent variables within the empirical models, in their current form, are correlated (multicollinear), and 2) current algorithms are global approaches and are based on the spatial stationarity assumption, so they are independent of location. Multicollinearity problem is resolved by using partial least squares (PLS). PLS, which transforms the data into a set of independent components, can be considered as a combined form of principal component regression (PCR) and multiple regression. Geographically weighted regression (GWR) is also used to investigate the validity of spatial stationarity assumption. GWR solves a regression model over each sample point by using the observations within its neighbourhood. PLS results show that the empirical method underestimates chlorophyll content in high latitudes, including the Southern Ocean region, when compared to PLS (see Figure 1). Cluster analysis of GWR coefficients also shows that the spatial stationarity assumption in empirical models is not likely a valid assumption.

Determination of cellulose I crystallinity by FT-Raman spectroscopy

Treesearch

Umesh P. Agarwal; Richard S. Reiner; Sally A. Ralph

2009-01-01

Two new methods based on FT-Raman spectroscopy, one simple, based on band intensity ratio, and the other, using a partial least-squares (PLS) regression model, are proposed to determine cellulose I crystallinity. In the simple method, crystallinity in semicrystalline cellulose I samples was determined based on univariate regression that was first developed using the...

CONTRIBUTION OF NUTRIENTS AND E. COLI TO SURFACE WATER CONDITION IN THE OZARKS I. USING PARTIAL LEAST SQUARES PREDICTIONS WHEN STANDARD REGRESSION ASSUMPTIONS ARE VIOLATED

EPA Science Inventory

We present here the application of PLS regression to predicting surface water total phosphorous, total ammonia and Escherichia coli from landscape metrics. The amount of variability in surface water constituents explained by each model reflects the composition of the contributi...

Image interpolation via regularized local linear regression.

PubMed

Liu, Xianming; Zhao, Debin; Xiong, Ruiqin; Ma, Siwei; Gao, Wen; Sun, Huifang

2011-12-01

The linear regression model is a very attractive tool to design effective image interpolation schemes. Some regression-based image interpolation algorithms have been proposed in the literature, in which the objective functions are optimized by ordinary least squares (OLS). However, it is shown that interpolation with OLS may have some undesirable properties from a robustness point of view: even small amounts of outliers can dramatically affect the estimates. To address these issues, in this paper we propose a novel image interpolation algorithm based on regularized local linear regression (RLLR). Starting with the linear regression model where we replace the OLS error norm with the moving least squares (MLS) error norm leads to a robust estimator of local image structure. To keep the solution stable and avoid overfitting, we incorporate the l(2)-norm as the estimator complexity penalty. Moreover, motivated by recent progress on manifold-based semi-supervised learning, we explicitly consider the intrinsic manifold structure by making use of both measured and unmeasured data points. Specifically, our framework incorporates the geometric structure of the marginal probability distribution induced by unmeasured samples as an additional local smoothness preserving constraint. The optimal model parameters can be obtained with a closed-form solution by solving a convex optimization problem. Experimental results on benchmark test images demonstrate that the proposed method achieves very competitive performance with the state-of-the-art interpolation algorithms, especially in image edge structure preservation. © 2011 IEEE

Bias correction by use of errors-in-variables regression models in studies with K-X-ray fluorescence bone lead measurements.

PubMed

Lamadrid-Figueroa, Héctor; Téllez-Rojo, Martha M; Angeles, Gustavo; Hernández-Ávila, Mauricio; Hu, Howard

2011-01-01

In-vivo measurement of bone lead by means of K-X-ray fluorescence (KXRF) is the preferred biological marker of chronic exposure to lead. Unfortunately, considerable measurement error associated with KXRF estimations can introduce bias in estimates of the effect of bone lead when this variable is included as the exposure in a regression model. Estimates of uncertainty reported by the KXRF instrument reflect the variance of the measurement error and, although they can be used to correct the measurement error bias, they are seldom used in epidemiological statistical analyzes. Errors-in-variables regression (EIV) allows for correction of bias caused by measurement error in predictor variables, based on the knowledge of the reliability of such variables. The authors propose a way to obtain reliability coefficients for bone lead measurements from uncertainty data reported by the KXRF instrument and compare, by the use of Monte Carlo simulations, results obtained using EIV regression models vs. those obtained by the standard procedures. Results of the simulations show that Ordinary Least Square (OLS) regression models provide severely biased estimates of effect, and that EIV provides nearly unbiased estimates. Although EIV effect estimates are more imprecise, their mean squared error is much smaller than that of OLS estimates. In conclusion, EIV is a better alternative than OLS to estimate the effect of bone lead when measured by KXRF. Copyright © 2010 Elsevier Inc. All rights reserved.

An Analysis of Advertising Effectiveness for U.S. Navy Recruiting

DTIC Science & Technology

1997-09-01

This thesis estimates the effect of Navy television advertising on enlistment rates of high quality male recruits (Armed Forces Qualification Test...Joint advertising is for all Armed Forces), Joint journal, and Joint direct mail advertising are explored. Enlistments are modeled as a function of...several factors including advertising , recruiters, and economic. Regression analyses (Ordinary Least Squares and Two Stage Least Squares) explore the

Partial Least Squares Regression Calibration of an Ultraviolet-Visible Spectrophotometer for Measurements of Chemical Oxygen Demand in Dye Wastewater

NASA Astrophysics Data System (ADS)

Mai, W.; Zhang, J.-F.; Zhao, X.-M.; Li, Z.; Xu, Z.-W.

2017-11-01

Wastewater from the dye industry is typically analyzed using a standard method for measurement of chemical oxygen demand (COD) or by a single-wavelength spectroscopic method. To overcome the disadvantages of these methods, ultraviolet-visible (UV-Vis) spectroscopy was combined with principal component regression (PCR) and partial least squares regression (PLSR) in this study. Unlike the standard method, this method does not require digestion of the samples for preparation. Experiments showed that the PLSR model offered high prediction performance for COD, with a mean relative error of about 5% for two dyes. This error is similar to that obtained with the standard method. In this study, the precision of the PLSR model decreased with the number of dye compounds present. It is likely that multiple models will be required in reality, and the complexity of a COD monitoring system would be greatly reduced if the PLSR model is used because it can include several dyes. UV-Vis spectroscopy with PLSR successfully enhanced the performance of COD prediction for dye wastewater and showed good potential for application in on-line water quality monitoring.

Effects of Employing Ridge Regression in Structural Equation Models.

ERIC Educational Resources Information Center

McQuitty, Shaun

1997-01-01

LISREL 8 invokes a ridge option when maximum likelihood or generalized least squares are used to estimate a structural equation model with a nonpositive definite covariance or correlation matrix. Implications of the ridge option for model fit, parameter estimates, and standard errors are explored through two examples. (SLD)

Regression model development and computational procedures to support estimation of real-time concentrations and loads of selected constituents in two tributaries to Lake Houston near Houston, Texas, 2005-9

USGS Publications Warehouse

Lee, Michael T.; Asquith, William H.; Oden, Timothy D.

2012-01-01

In December 2005, the U.S. Geological Survey (USGS), in cooperation with the City of Houston, Texas, began collecting discrete water-quality samples for nutrients, total organic carbon, bacteria (Escherichia coli and total coliform), atrazine, and suspended sediment at two USGS streamflow-gaging stations that represent watersheds contributing to Lake Houston (08068500 Spring Creek near Spring, Tex., and 08070200 East Fork San Jacinto River near New Caney, Tex.). Data from the discrete water-quality samples collected during 2005–9, in conjunction with continuously monitored real-time data that included streamflow and other physical water-quality properties (specific conductance, pH, water temperature, turbidity, and dissolved oxygen), were used to develop regression models for the estimation of concentrations of water-quality constituents of substantial source watersheds to Lake Houston. The potential explanatory variables included discharge (streamflow), specific conductance, pH, water temperature, turbidity, dissolved oxygen, and time (to account for seasonal variations inherent in some water-quality data). The response variables (the selected constituents) at each site were nitrite plus nitrate nitrogen, total phosphorus, total organic carbon, E. coli, atrazine, and suspended sediment. The explanatory variables provide easily measured quantities to serve as potential surrogate variables to estimate concentrations of the selected constituents through statistical regression. Statistical regression also facilitates accompanying estimates of uncertainty in the form of prediction intervals. Each regression model potentially can be used to estimate concentrations of a given constituent in real time. Among other regression diagnostics, the diagnostics used as indicators of general model reliability and reported herein include the adjusted R-squared, the residual standard error, residual plots, and p-values. Adjusted R-squared values for the Spring Creek models ranged from .582–.922 (dimensionless). The residual standard errors ranged from .073–.447 (base-10 logarithm). Adjusted R-squared values for the East Fork San Jacinto River models ranged from .253–.853 (dimensionless). The residual standard errors ranged from .076–.388 (base-10 logarithm). In conjunction with estimated concentrations, constituent loads can be estimated by multiplying the estimated concentration by the corresponding streamflow and by applying the appropriate conversion factor. The regression models presented in this report are site specific, that is, they are specific to the Spring Creek and East Fork San Jacinto River streamflow-gaging stations; however, the general methods that were developed and documented could be applied to most perennial streams for the purpose of estimating real-time water quality data.

Wavelet regression model in forecasting crude oil price

NASA Astrophysics Data System (ADS)

Hamid, Mohd Helmie; Shabri, Ani

2017-05-01

This study presents the performance of wavelet multiple linear regression (WMLR) technique in daily crude oil forecasting. WMLR model was developed by integrating the discrete wavelet transform (DWT) and multiple linear regression (MLR) model. The original time series was decomposed to sub-time series with different scales by wavelet theory. Correlation analysis was conducted to assist in the selection of optimal decomposed components as inputs for the WMLR model. The daily WTI crude oil price series has been used in this study to test the prediction capability of the proposed model. The forecasting performance of WMLR model were also compared with regular multiple linear regression (MLR), Autoregressive Moving Average (ARIMA) and Generalized Autoregressive Conditional Heteroscedasticity (GARCH) using root mean square errors (RMSE) and mean absolute errors (MAE). Based on the experimental results, it appears that the WMLR model performs better than the other forecasting technique tested in this study.

Improved accuracy in quantitative laser-induced breakdown spectroscopy using sub-models

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

Anderson, Ryan B.; Clegg, Samuel M.; Frydenvang, Jens

We report that accurate quantitative analysis of diverse geologic materials is one of the primary challenges faced by the Laser-Induced Breakdown Spectroscopy (LIBS)-based ChemCam instrument on the Mars Science Laboratory (MSL) rover. The SuperCam instrument on the Mars 2020 rover, as well as other LIBS instruments developed for geochemical analysis on Earth or other planets, will face the same challenge. Consequently, part of the ChemCam science team has focused on the development of improved multivariate analysis calibrations methods. Developing a single regression model capable of accurately determining the composition of very different target materials is difficult because the response ofmore » an element’s emission lines in LIBS spectra can vary with the concentration of other elements. We demonstrate a conceptually simple “submodel” method for improving the accuracy of quantitative LIBS analysis of diverse target materials. The method is based on training several regression models on sets of targets with limited composition ranges and then “blending” these “sub-models” into a single final result. Tests of the sub-model method show improvement in test set root mean squared error of prediction (RMSEP) for almost all cases. Lastly, the sub-model method, using partial least squares regression (PLS), is being used as part of the current ChemCam quantitative calibration, but the sub-model method is applicable to any multivariate regression method and may yield similar improvements.« less

Improved accuracy in quantitative laser-induced breakdown spectroscopy using sub-models

DOE PAGES

Anderson, Ryan B.; Clegg, Samuel M.; Frydenvang, Jens; ...

2016-12-15

We report that accurate quantitative analysis of diverse geologic materials is one of the primary challenges faced by the Laser-Induced Breakdown Spectroscopy (LIBS)-based ChemCam instrument on the Mars Science Laboratory (MSL) rover. The SuperCam instrument on the Mars 2020 rover, as well as other LIBS instruments developed for geochemical analysis on Earth or other planets, will face the same challenge. Consequently, part of the ChemCam science team has focused on the development of improved multivariate analysis calibrations methods. Developing a single regression model capable of accurately determining the composition of very different target materials is difficult because the response ofmore » an element’s emission lines in LIBS spectra can vary with the concentration of other elements. We demonstrate a conceptually simple “submodel” method for improving the accuracy of quantitative LIBS analysis of diverse target materials. The method is based on training several regression models on sets of targets with limited composition ranges and then “blending” these “sub-models” into a single final result. Tests of the sub-model method show improvement in test set root mean squared error of prediction (RMSEP) for almost all cases. Lastly, the sub-model method, using partial least squares regression (PLS), is being used as part of the current ChemCam quantitative calibration, but the sub-model method is applicable to any multivariate regression method and may yield similar improvements.« less

Robust scoring functions for protein-ligand interactions with quantum chemical charge models.

PubMed

Wang, Jui-Chih; Lin, Jung-Hsin; Chen, Chung-Ming; Perryman, Alex L; Olson, Arthur J

2011-10-24

Ordinary least-squares (OLS) regression has been used widely for constructing the scoring functions for protein-ligand interactions. However, OLS is very sensitive to the existence of outliers, and models constructed using it are easily affected by the outliers or even the choice of the data set. On the other hand, determination of atomic charges is regarded as of central importance, because the electrostatic interaction is known to be a key contributing factor for biomolecular association. In the development of the AutoDock4 scoring function, only OLS was conducted, and the simple Gasteiger method was adopted. It is therefore of considerable interest to see whether more rigorous charge models could improve the statistical performance of the AutoDock4 scoring function. In this study, we have employed two well-established quantum chemical approaches, namely the restrained electrostatic potential (RESP) and the Austin-model 1-bond charge correction (AM1-BCC) methods, to obtain atomic partial charges, and we have compared how different charge models affect the performance of AutoDock4 scoring functions. In combination with robust regression analysis and outlier exclusion, our new protein-ligand free energy regression model with AM1-BCC charges for ligands and Amber99SB charges for proteins achieve lowest root-mean-squared error of 1.637 kcal/mol for the training set of 147 complexes and 2.176 kcal/mol for the external test set of 1427 complexes. The assessment for binding pose prediction with the 100 external decoy sets indicates very high success rate of 87% with the criteria of predicted root-mean-squared deviation of less than 2 Å. The success rates and statistical performance of our robust scoring functions are only weakly class-dependent (hydrophobic, hydrophilic, or mixed).

Analysis of low flows and selected methods for estimating low-flow characteristics at partial-record and ungaged stream sites in western Washington

USGS Publications Warehouse

Curran, Christopher A.; Eng, Ken; Konrad, Christopher P.

2012-01-01

Regional low-flow regression models for estimating Q7,10 at ungaged stream sites are developed from the records of daily discharge at 65 continuous gaging stations (including 22 discontinued gaging stations) for the purpose of evaluating explanatory variables. By incorporating the base-flow recession time constant τ as an explanatory variable in the regression model, the root-mean square error for estimating Q7,10 at ungaged sites can be lowered to 72 percent (for known values of τ), which is 42 percent less than if only basin area and mean annual precipitation are used as explanatory variables. If partial-record sites are included in the regression data set, τ must be estimated from pairs of discharge measurements made during continuous periods of declining low flows. Eight measurement pairs are optimal for estimating τ at partial-record sites, and result in a lowering of the root-mean square error by 25 percent. A low-flow survey strategy that includes paired measurements at partial-record sites requires additional effort and planning beyond a standard strategy, but could be used to enhance regional estimates of τ and potentially reduce the error of regional regression models for estimating low-flow characteristics at ungaged sites.

Solid-phase cadmium speciation in soil using L3-edge XANES spectroscopy with partial least-squares regression.

PubMed

Siebers, Nina; Kruse, Jens; Eckhardt, Kai-Uwe; Hu, Yongfeng; Leinweber, Peter

2012-07-01

Cadmium (Cd) has a high toxicity and resolving its speciation in soil is challenging but essential for estimating the environmental risk. In this study partial least-square (PLS) regression was tested for its capability to deconvolute Cd L(3)-edge X-ray absorption near-edge structure (XANES) spectra of multi-compound mixtures. For this, a library of Cd reference compound spectra and a spectrum of a soil sample were acquired. A good coefficient of determination (R(2)) of Cd compounds in mixtures was obtained for the PLS model using binary and ternary mixtures of various Cd reference compounds proving the validity of this approach. In order to describe complex systems like soil, multi-compound mixtures of a variety of Cd compounds must be included in the PLS model. The obtained PLS regression model was then applied to a highly Cd-contaminated soil revealing Cd(3)(PO(4))(2) (36.1%), Cd(NO(3))(2)·4H(2)O (24.5%), Cd(OH)(2) (21.7%), CdCO(3) (17.1%) and CdCl(2) (0.4%). These preliminary results proved that PLS regression is a promising approach for a direct determination of Cd speciation in the solid phase of a soil sample.

Developing a dengue forecast model using machine learning: A case study in China

PubMed Central

Zhang, Qin; Wang, Li; Xiao, Jianpeng; Zhang, Qingying; Luo, Ganfeng; Li, Zhihao; He, Jianfeng; Zhang, Yonghui; Ma, Wenjun

2017-01-01

Background In China, dengue remains an important public health issue with expanded areas and increased incidence recently. Accurate and timely forecasts of dengue incidence in China are still lacking. We aimed to use the state-of-the-art machine learning algorithms to develop an accurate predictive model of dengue. Methodology/Principal findings Weekly dengue cases, Baidu search queries and climate factors (mean temperature, relative humidity and rainfall) during 2011–2014 in Guangdong were gathered. A dengue search index was constructed for developing the predictive models in combination with climate factors. The observed year and week were also included in the models to control for the long-term trend and seasonality. Several machine learning algorithms, including the support vector regression (SVR) algorithm, step-down linear regression model, gradient boosted regression tree algorithm (GBM), negative binomial regression model (NBM), least absolute shrinkage and selection operator (LASSO) linear regression model and generalized additive model (GAM), were used as candidate models to predict dengue incidence. Performance and goodness of fit of the models were assessed using the root-mean-square error (RMSE) and R-squared measures. The residuals of the models were examined using the autocorrelation and partial autocorrelation function analyses to check the validity of the models. The models were further validated using dengue surveillance data from five other provinces. The epidemics during the last 12 weeks and the peak of the 2014 large outbreak were accurately forecasted by the SVR model selected by a cross-validation technique. Moreover, the SVR model had the consistently smallest prediction error rates for tracking the dynamics of dengue and forecasting the outbreaks in other areas in China. Conclusion and significance The proposed SVR model achieved a superior performance in comparison with other forecasting techniques assessed in this study. The findings can help the government and community respond early to dengue epidemics. PMID:29036169

Determination of total iron-reactive phenolics, anthocyanins and tannins in wine grapes of skins and seeds based on near-infrared hyperspectral imaging.

PubMed

Zhang, Ni; Liu, Xu; Jin, Xiaoduo; Li, Chen; Wu, Xuan; Yang, Shuqin; Ning, Jifeng; Yanne, Paul

2017-12-15

Phenolics contents in wine grapes are key indicators for assessing ripeness. Near-infrared hyperspectral images during ripening have been explored to achieve an effective method for predicting phenolics contents. Principal component regression (PCR), partial least squares regression (PLSR) and support vector regression (SVR) models were built, respectively. The results show that SVR behaves globally better than PLSR and PCR, except in predicting tannins content of seeds. For the best prediction results, the squared correlation coefficient and root mean square error reached 0.8960 and 0.1069g/L (+)-catechin equivalents (CE), respectively, for tannins in skins, 0.9065 and 0.1776 (g/L CE) for total iron-reactive phenolics (TIRP) in skins, 0.8789 and 0.1442 (g/L M3G) for anthocyanins in skins, 0.9243 and 0.2401 (g/L CE) for tannins in seeds, and 0.8790 and 0.5190 (g/L CE) for TIRP in seeds. Our results indicated that NIR hyperspectral imaging has good prospects for evaluation of phenolics in wine grapes. Copyright © 2017 Elsevier Ltd. All rights reserved.

The relationship between air pollution, fossil fuel energy consumption, and water resources in the panel of selected Asia-Pacific countries.

PubMed

Rafindadi, Abdulkadir Abdulrashid; Yusof, Zarinah; Zaman, Khalid; Kyophilavong, Phouphet; Akhmat, Ghulam

2014-10-01

The objective of the study is to examine the relationship between air pollution, fossil fuel energy consumption, water resources, and natural resource rents in the panel of selected Asia-Pacific countries, over a period of 1975-2012. The study includes number of variables in the model for robust analysis. The results of cross-sectional analysis show that there is a significant relationship between air pollution, energy consumption, and water productivity in the individual countries of Asia-Pacific. However, the results of each country vary according to the time invariant shocks. For this purpose, the study employed the panel least square technique which includes the panel least square regression, panel fixed effect regression, and panel two-stage least square regression. In general, all the panel tests indicate that there is a significant and positive relationship between air pollution, energy consumption, and water resources in the region. The fossil fuel energy consumption has a major dominating impact on the changes in the air pollution in the region.

Quantitative Structure Retention Relationships of Polychlorinated Dibenzodioxins and Dibenzofurans

DTIC Science & Technology

1991-08-01

be a projection onto the X-Y plane. The algorithm for this calculation can be found in Stouch and Jurs (22), but was further refined by Rohrbaugh and...throughspace distances. WPSA2 (c) Weighted positive charged surface area. MOMH2 (c) Second major moment of inertia with hydrogens attached. CSTR 3 (d) Sum...of the models. The robust regression analysis method calculates a regression model using a least median squares algorithm which is not as susceptible

Raman spectroscopy based investigation of molecular changes associated with an early stage of dengue virus infection

NASA Astrophysics Data System (ADS)

Bilal, Maria; Bilal, Muhammad; Saleem, Muhammad; Khurram, Muhammad; Khan, Saranjam; Ullah, Rahat; Ali, Hina; Ahmed, Mushtaq; Shahzada, Shaista; Ullah Khan, Ehsan

2017-04-01

Raman spectroscopy based investigations of the molecular changes associated with an early stage of dengue virus infection (DENV) using a partial least squares (PLS) regression model is presented. This study is based on non-structural protein 1 (NS1) which appears after three days of DENV infection. In total, 39 blood sera samples were collected and divided into two groups. The control group contained samples which were the negative for NS1 and antibodies and the positive group contained those samples in which NS1 is positive and antibodies were negative. Out of 39 samples, 29 Raman spectra were used for the model development while the remaining 10 were kept hidden for blind testing of the model. PLS regression yielded a vector of regression coefficients as a function of Raman shift, which were analyzed. Cytokines in the region 775-875 cm-1, lectins at 1003, 1238, 1340, 1449 and 1672 cm-1, DNA in the region 1040-1140 cm-1 and alpha and beta structures of proteins in the region 933-967 cm-1 have been identified in the regression vector for their role in an early stage of DENV infection. Validity of the model was established by its R-square value of 0.891. Sensitivity, specificity and accuracy were 100% each and the area under the receiver operator characteristic curve was found to be 1.

Bivariate least squares linear regression: Towards a unified analytic formalism. I. Functional models

NASA Astrophysics Data System (ADS)

Caimmi, R.

2011-08-01

Concerning bivariate least squares linear regression, the classical approach pursued for functional models in earlier attempts ( York, 1966, 1969) is reviewed using a new formalism in terms of deviation (matrix) traces which, for unweighted data, reduce to usual quantities leaving aside an unessential (but dimensional) multiplicative factor. Within the framework of classical error models, the dependent variable relates to the independent variable according to the usual additive model. The classes of linear models considered are regression lines in the general case of correlated errors in X and in Y for weighted data, and in the opposite limiting situations of (i) uncorrelated errors in X and in Y, and (ii) completely correlated errors in X and in Y. The special case of (C) generalized orthogonal regression is considered in detail together with well known subcases, namely: (Y) errors in X negligible (ideally null) with respect to errors in Y; (X) errors in Y negligible (ideally null) with respect to errors in X; (O) genuine orthogonal regression; (R) reduced major-axis regression. In the limit of unweighted data, the results determined for functional models are compared with their counterparts related to extreme structural models i.e. the instrumental scatter is negligible (ideally null) with respect to the intrinsic scatter ( Isobe et al., 1990; Feigelson and Babu, 1992). While regression line slope and intercept estimators for functional and structural models necessarily coincide, the contrary holds for related variance estimators even if the residuals obey a Gaussian distribution, with the exception of Y models. An example of astronomical application is considered, concerning the [O/H]-[Fe/H] empirical relations deduced from five samples related to different stars and/or different methods of oxygen abundance determination. For selected samples and assigned methods, different regression models yield consistent results within the errors (∓ σ) for both heteroscedastic and homoscedastic data. Conversely, samples related to different methods produce discrepant results, due to the presence of (still undetected) systematic errors, which implies no definitive statement can be made at present. A comparison is also made between different expressions of regression line slope and intercept variance estimators, where fractional discrepancies are found to be not exceeding a few percent, which grows up to about 20% in the presence of large dispersion data. An extension of the formalism to structural models is left to a forthcoming paper.

Linear regression based on Minimum Covariance Determinant (MCD) and TELBS methods on the productivity of phytoplankton

NASA Astrophysics Data System (ADS)

Gusriani, N.; Firdaniza

2018-03-01

The existence of outliers on multiple linear regression analysis causes the Gaussian assumption to be unfulfilled. If the Least Square method is forcedly used on these data, it will produce a model that cannot represent most data. For that, we need a robust regression method against outliers. This paper will compare the Minimum Covariance Determinant (MCD) method and the TELBS method on secondary data on the productivity of phytoplankton, which contains outliers. Based on the robust determinant coefficient value, MCD method produces a better model compared to TELBS method.

An Empirical Model Building Criterion Based on Prediction with Applications in Parametric Cost Estimation.

DTIC Science & Technology

1980-08-01

varia- ble is denoted by 7, the total sum of squares of deviations from that mean is defined by n - SSTO - (-Y) (2.6) iul and the regression sum of...squares by SSR - SSTO - SSE (2.7) II 14 A selection criterion is a rule according to which a certain model out of the 2p possible models is labeled "best...dis- cussed next. 1. The R2 Criterion The coefficient of determination is defined by R2 . 1 - SSE/ SSTO . (2.8) It is clear that R is the proportion of

Efficient design of gain-flattened multi-pump Raman fiber amplifiers using least squares support vector regression

NASA Astrophysics Data System (ADS)

Chen, Jing; Qiu, Xiaojie; Yin, Cunyi; Jiang, Hao

2018-02-01

An efficient method to design the broadband gain-flattened Raman fiber amplifier with multiple pumps is proposed based on least squares support vector regression (LS-SVR). A multi-input multi-output LS-SVR model is introduced to replace the complicated solving process of the nonlinear coupled Raman amplification equation. The proposed approach contains two stages: offline training stage and online optimization stage. During the offline stage, the LS-SVR model is trained. Owing to the good generalization capability of LS-SVR, the net gain spectrum can be directly and accurately obtained when inputting any combination of the pump wavelength and power to the well-trained model. During the online stage, we incorporate the LS-SVR model into the particle swarm optimization algorithm to find the optimal pump configuration. The design results demonstrate that the proposed method greatly shortens the computation time and enhances the efficiency of the pump parameter optimization for Raman fiber amplifier design.

Near infra red spectroscopy as a multivariate process analytical tool for predicting pharmaceutical co-crystal concentration.

PubMed

Wood, Clive; Alwati, Abdolati; Halsey, Sheelagh; Gough, Tim; Brown, Elaine; Kelly, Adrian; Paradkar, Anant

2016-09-10

The use of near infra red spectroscopy to predict the concentration of two pharmaceutical co-crystals; 1:1 ibuprofen-nicotinamide (IBU-NIC) and 1:1 carbamazepine-nicotinamide (CBZ-NIC) has been evaluated. A partial least squares (PLS) regression model was developed for both co-crystal pairs using sets of standard samples to create calibration and validation data sets with which to build and validate the models. Parameters such as the root mean square error of calibration (RMSEC), root mean square error of prediction (RMSEP) and correlation coefficient were used to assess the accuracy and linearity of the models. Accurate PLS regression models were created for both co-crystal pairs which can be used to predict the co-crystal concentration in a powder mixture of the co-crystal and the active pharmaceutical ingredient (API). The IBU-NIC model had smaller errors than the CBZ-NIC model, possibly due to the complex CBZ-NIC spectra which could reflect the different arrangement of hydrogen bonding associated with the co-crystal compared to the IBU-NIC co-crystal. These results suggest that NIR spectroscopy can be used as a PAT tool during a variety of pharmaceutical co-crystal manufacturing methods and the presented data will facilitate future offline and in-line NIR studies involving pharmaceutical co-crystals. Copyright © 2016 The Authors. Published by Elsevier B.V. All rights reserved.

An analysis of input errors in precipitation-runoff models using regression with errors in the independent variables

USGS Publications Warehouse

Troutman, Brent M.

1982-01-01

Errors in runoff prediction caused by input data errors are analyzed by treating precipitation-runoff models as regression (conditional expectation) models. Independent variables of the regression consist of precipitation and other input measurements; the dependent variable is runoff. In models using erroneous input data, prediction errors are inflated and estimates of expected storm runoff for given observed input variables are biased. This bias in expected runoff estimation results in biased parameter estimates if these parameter estimates are obtained by a least squares fit of predicted to observed runoff values. The problems of error inflation and bias are examined in detail for a simple linear regression of runoff on rainfall and for a nonlinear U.S. Geological Survey precipitation-runoff model. Some implications for flood frequency analysis are considered. A case study using a set of data from Turtle Creek near Dallas, Texas illustrates the problems of model input errors.

On the calibration process of film dosimetry: OLS inverse regression versus WLS inverse prediction.

PubMed

Crop, F; Van Rompaye, B; Paelinck, L; Vakaet, L; Thierens, H; De Wagter, C

2008-07-21

The purpose of this study was both putting forward a statistically correct model for film calibration and the optimization of this process. A reliable calibration is needed in order to perform accurate reference dosimetry with radiographic (Gafchromic) film. Sometimes, an ordinary least squares simple linear (in the parameters) regression is applied to the dose-optical-density (OD) curve with the dose as a function of OD (inverse regression) or sometimes OD as a function of dose (inverse prediction). The application of a simple linear regression fit is an invalid method because heteroscedasticity of the data is not taken into account. This could lead to erroneous results originating from the calibration process itself and thus to a lower accuracy. In this work, we compare the ordinary least squares (OLS) inverse regression method with the correct weighted least squares (WLS) inverse prediction method to create calibration curves. We found that the OLS inverse regression method could lead to a prediction bias of up to 7.3 cGy at 300 cGy and total prediction errors of 3% or more for Gafchromic EBT film. Application of the WLS inverse prediction method resulted in a maximum prediction bias of 1.4 cGy and total prediction errors below 2% in a 0-400 cGy range. We developed a Monte-Carlo-based process to optimize calibrations, depending on the needs of the experiment. This type of thorough analysis can lead to a higher accuracy for film dosimetry.

Determinants of Antibiotic Consumption - Development of a Model using Partial Least Squares Regression based on Data from India.

PubMed

Tamhankar, Ashok J; Karnik, Shreyasee S; Stålsby Lundborg, Cecilia

2018-04-23

Antibiotic resistance, a consequence of antibiotic use, is a threat to health, with severe consequences for resource constrained settings. If determinants for human antibiotic use in India, a lower middle income country, with one of the highest antibiotic consumption in the world could be understood, interventions could be developed, having implications for similar settings. Year wise data for India, for potential determinants and antibiotic consumption, was sourced from publicly available databases for the years 2000-2010. Data was analyzed using Partial Least Squares regression and correlation between determinants and antibiotic consumption was evaluated, formulating 'Predictors' and 'Prediction models'. The 'prediction model' with the statistically most significant predictors (root mean square errors of prediction for train set-377.0 and test set-297.0) formulated from a combination of Health infrastructure + Surface transport infrastructure (HISTI), predicted antibiotic consumption within 95% confidence interval and estimated an antibiotic consumption of 11.6 standard units/person (14.37 billion standard units totally; standard units = number of doses sold in the country; a dose being a pill, capsule, or ampoule) for India for 2014. The HISTI model may become useful in predicting antibiotic consumption for countries/regions having circumstances and data similar to India, but without resources to measure actual data of antibiotic consumption.

Moderation analysis using a two-level regression model.

PubMed

Yuan, Ke-Hai; Cheng, Ying; Maxwell, Scott

2014-10-01

Moderation analysis is widely used in social and behavioral research. The most commonly used model for moderation analysis is moderated multiple regression (MMR) in which the explanatory variables of the regression model include product terms, and the model is typically estimated by least squares (LS). This paper argues for a two-level regression model in which the regression coefficients of a criterion variable on predictors are further regressed on moderator variables. An algorithm for estimating the parameters of the two-level model by normal-distribution-based maximum likelihood (NML) is developed. Formulas for the standard errors (SEs) of the parameter estimates are provided and studied. Results indicate that, when heteroscedasticity exists, NML with the two-level model gives more efficient and more accurate parameter estimates than the LS analysis of the MMR model. When error variances are homoscedastic, NML with the two-level model leads to essentially the same results as LS with the MMR model. Most importantly, the two-level regression model permits estimating the percentage of variance of each regression coefficient that is due to moderator variables. When applied to data from General Social Surveys 1991, NML with the two-level model identified a significant moderation effect of race on the regression of job prestige on years of education while LS with the MMR model did not. An R package is also developed and documented to facilitate the application of the two-level model.

Cellulose I crystallinity determination using FT-Raman spectroscopy : univariate and multivariate methods

Treesearch

Umesh P. Agarwal; Richard S. Reiner; Sally A. Ralph

2010-01-01

Two new methods based on FT–Raman spectroscopy, one simple, based on band intensity ratio, and the other using a partial least squares (PLS) regression model, are proposed to determine cellulose I crystallinity. In the simple method, crystallinity in cellulose I samples was determined based on univariate regression that was first developed using the Raman band...

Avoiding and Correcting Bias in Score-Based Latent Variable Regression with Discrete Manifest Items

ERIC Educational Resources Information Center

Lu, Irene R. R.; Thomas, D. Roland

2008-01-01

This article considers models involving a single structural equation with latent explanatory and/or latent dependent variables where discrete items are used to measure the latent variables. Our primary focus is the use of scores as proxies for the latent variables and carrying out ordinary least squares (OLS) regression on such scores to estimate…

Quantitative analysis of binary polymorphs mixtures of fusidic acid by diffuse reflectance FTIR spectroscopy, diffuse reflectance FT-NIR spectroscopy, Raman spectroscopy and multivariate calibration.

PubMed

Guo, Canyong; Luo, Xuefang; Zhou, Xiaohua; Shi, Beijia; Wang, Juanjuan; Zhao, Jinqi; Zhang, Xiaoxia

2017-06-05

Vibrational spectroscopic techniques such as infrared, near-infrared and Raman spectroscopy have become popular in detecting and quantifying polymorphism of pharmaceutics since they are fast and non-destructive. This study assessed the ability of three vibrational spectroscopy combined with multivariate analysis to quantify a low-content undesired polymorph within a binary polymorphic mixture. Partial least squares (PLS) regression and support vector machine (SVM) regression were employed to build quantitative models. Fusidic acid, a steroidal antibiotic, was used as the model compound. It was found that PLS regression performed slightly better than SVM regression in all the three spectroscopic techniques. Root mean square errors of prediction (RMSEP) were ranging from 0.48% to 1.17% for diffuse reflectance FTIR spectroscopy and 1.60-1.93% for diffuse reflectance FT-NIR spectroscopy and 1.62-2.31% for Raman spectroscopy. The results indicate that diffuse reflectance FTIR spectroscopy offers significant advantages in providing accurate measurement of polymorphic content in the fusidic acid binary mixtures, while Raman spectroscopy is the least accurate technique for quantitative analysis of polymorphs. Copyright © 2017 Elsevier B.V. All rights reserved.

A Simple Introduction to Moving Least Squares and Local Regression Estimation

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

Garimella, Rao Veerabhadra

In this brief note, a highly simpli ed introduction to esimating functions over a set of particles is presented. The note starts from Global Least Squares tting, going on to Moving Least Squares estimation (MLS) and nally, Local Regression Estimation (LRE).

Prediction of pH of cola beverage using Vis/NIR spectroscopy and least squares-support vector machine

NASA Astrophysics Data System (ADS)

Liu, Fei; He, Yong

2008-02-01

Visible and near infrared (Vis/NIR) transmission spectroscopy and chemometric methods were utilized to predict the pH values of cola beverages. Five varieties of cola were prepared and 225 samples (45 samples for each variety) were selected for the calibration set, while 75 samples (15 samples for each variety) for the validation set. The smoothing way of Savitzky-Golay and standard normal variate (SNV) followed by first-derivative were used as the pre-processing methods. Partial least squares (PLS) analysis was employed to extract the principal components (PCs) which were used as the inputs of least squares-support vector machine (LS-SVM) model according to their accumulative reliabilities. Then LS-SVM with radial basis function (RBF) kernel function and a two-step grid search technique were applied to build the regression model with a comparison of PLS regression. The correlation coefficient (r), root mean square error of prediction (RMSEP) and bias were 0.961, 0.040 and 0.012 for PLS, while 0.975, 0.031 and 4.697x10 -3 for LS-SVM, respectively. Both methods obtained a satisfying precision. The results indicated that Vis/NIR spectroscopy combined with chemometric methods could be applied as an alternative way for the prediction of pH of cola beverages.

Domain-Invariant Partial-Least-Squares Regression.

PubMed

Nikzad-Langerodi, Ramin; Zellinger, Werner; Lughofer, Edwin; Saminger-Platz, Susanne

2018-05-11

Multivariate calibration models often fail to extrapolate beyond the calibration samples because of changes associated with the instrumental response, environmental condition, or sample matrix. Most of the current methods used to adapt a source calibration model to a target domain exclusively apply to calibration transfer between similar analytical devices, while generic methods for calibration-model adaptation are largely missing. To fill this gap, we here introduce domain-invariant partial-least-squares (di-PLS) regression, which extends ordinary PLS by a domain regularizer in order to align the source and target distributions in the latent-variable space. We show that a domain-invariant weight vector can be derived in closed form, which allows the integration of (partially) labeled data from the source and target domains as well as entirely unlabeled data from the latter. We test our approach on a simulated data set where the aim is to desensitize a source calibration model to an unknown interfering agent in the target domain (i.e., unsupervised model adaptation). In addition, we demonstrate unsupervised, semisupervised, and supervised model adaptation by di-PLS on two real-world near-infrared (NIR) spectroscopic data sets.

[Hyperspectral Estimation of Apple Tree Canopy LAI Based on SVM and RF Regression].

PubMed

Han, Zhao-ying; Zhu, Xi-cun; Fang, Xian-yi; Wang, Zhuo-yuan; Wang, Ling; Zhao, Geng-Xing; Jiang, Yuan-mao

2016-03-01

Leaf area index (LAI) is the dynamic index of crop population size. Hyperspectral technology can be used to estimate apple canopy LAI rapidly and nondestructively. It can be provide a reference for monitoring the tree growing and yield estimation. The Red Fuji apple trees of full bearing fruit are the researching objects. Ninety apple trees canopies spectral reflectance and LAI values were measured by the ASD Fieldspec3 spectrometer and LAI-2200 in thirty orchards in constant two years in Qixia research area of Shandong Province. The optimal vegetation indices were selected by the method of correlation analysis of the original spectral reflectance and vegetation indices. The models of predicting the LAI were built with the multivariate regression analysis method of support vector machine (SVM) and random forest (RF). The new vegetation indices, GNDVI527, ND-VI676, RVI682, FD-NVI656 and GRVI517 and the previous two main vegetation indices, NDVI670 and NDVI705, are in accordance with LAI. In the RF regression model, the calibration set decision coefficient C-R2 of 0.920 and validation set decision coefficient V-R2 of 0.889 are higher than the SVM regression model by 0.045 and 0.033 respectively. The root mean square error of calibration set C-RMSE of 0.249, the root mean square error validation set V-RMSE of 0.236 are lower than that of the SVM regression model by 0.054 and 0.058 respectively. Relative analysis of calibrating error C-RPD and relative analysis of validation set V-RPD reached 3.363 and 2.520, 0.598 and 0.262, respectively, which were higher than the SVM regression model. The measured and predicted the scatterplot trend line slope of the calibration set and validation set C-S and V-S are close to 1. The estimation result of RF regression model is better than that of the SVM. RF regression model can be used to estimate the LAI of red Fuji apple trees in full fruit period.

Simplified large African carnivore density estimators from track indices.

PubMed

Winterbach, Christiaan W; Ferreira, Sam M; Funston, Paul J; Somers, Michael J

2016-01-01

The range, population size and trend of large carnivores are important parameters to assess their status globally and to plan conservation strategies. One can use linear models to assess population size and trends of large carnivores from track-based surveys on suitable substrates. The conventional approach of a linear model with intercept may not intercept at zero, but may fit the data better than linear model through the origin. We assess whether a linear regression through the origin is more appropriate than a linear regression with intercept to model large African carnivore densities and track indices. We did simple linear regression with intercept analysis and simple linear regression through the origin and used the confidence interval for ß in the linear model y  =  αx  + ß, Standard Error of Estimate, Mean Squares Residual and Akaike Information Criteria to evaluate the models. The Lion on Clay and Low Density on Sand models with intercept were not significant ( P  > 0.05). The other four models with intercept and the six models thorough origin were all significant ( P  < 0.05). The models using linear regression with intercept all included zero in the confidence interval for ß and the null hypothesis that ß = 0 could not be rejected. All models showed that the linear model through the origin provided a better fit than the linear model with intercept, as indicated by the Standard Error of Estimate and Mean Square Residuals. Akaike Information Criteria showed that linear models through the origin were better and that none of the linear models with intercept had substantial support. Our results showed that linear regression through the origin is justified over the more typical linear regression with intercept for all models we tested. A general model can be used to estimate large carnivore densities from track densities across species and study areas. The formula observed track density = 3.26 × carnivore density can be used to estimate densities of large African carnivores using track counts on sandy substrates in areas where carnivore densities are 0.27 carnivores/100 km 2 or higher. To improve the current models, we need independent data to validate the models and data to test for non-linear relationship between track indices and true density at low densities.

Assessing Interval Estimation Methods for Hill Model Parameters in a High-Throughput Screening Context (SOT)

EPA Science Inventory

The Hill model of concentration-response is ubiquitous in toxicology, perhaps because its parameters directly relate to biologically significant metrics of toxicity such as efficacy and potency. Point estimates of these parameters obtained through least squares regression or maxi...

Assessing Interval Estimation Methods for Hill Model Parameters in a High-Throughput Screening Context (IVIVE meeting)

EPA Science Inventory

The Hill model of concentration-response is ubiquitous in toxicology, perhaps because its parameters directly relate to biologically significant metrics of toxicity such as efficacy and potency. Point estimates of these parameters obtained through least squares regression or maxi...

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.

Quantum algorithm for linear regression

NASA Astrophysics Data System (ADS)

Wang, Guoming

2017-07-01

We present a quantum algorithm for fitting a linear regression model to a given data set using the least-squares approach. Differently from previous algorithms which yield a quantum state encoding the optimal parameters, our algorithm outputs these numbers in the classical form. So by running it once, one completely determines the fitted model and then can use it to make predictions on new data at little cost. Moreover, our algorithm works in the standard oracle model, and can handle data sets with nonsparse design matrices. It runs in time poly( log2(N ) ,d ,κ ,1 /ɛ ) , where N is the size of the data set, d is the number of adjustable parameters, κ is the condition number of the design matrix, and ɛ is the desired precision in the output. We also show that the polynomial dependence on d and κ is necessary. Thus, our algorithm cannot be significantly improved. Furthermore, we also give a quantum algorithm that estimates the quality of the least-squares fit (without computing its parameters explicitly). This algorithm runs faster than the one for finding this fit, and can be used to check whether the given data set qualifies for linear regression in the first place.

Estimation of Finger Joint Angles Based on Electromechanical Sensing of Wrist Shape.

PubMed

Kawaguchi, Junki; Yoshimoto, Shunsuke; Kuroda, Yoshihiro; Oshiro, Osamu

2017-09-01

An approach to finger motion capture that places fewer restrictions on the usage environment and actions of the user is an important research topic in biomechanics and human-computer interaction. We proposed a system that electrically detects finger motion from the associated deformation of the wrist and estimates the finger joint angles using multiple regression models. A wrist-mounted sensing device with 16 electrodes detects deformation of the wrist from changes in electrical contact resistance at the skin. In this study, we experimentally investigated the accuracy of finger joint angle estimation, the adequacy of two multiple regression models, and the resolution of the estimation of total finger joint angles. In experiments, both the finger joint angles and the system output voltage were recorded as subjects performed flexion/extension of the fingers. These data were used for calibration using the least-squares method. The system was found to be capable of estimating the total finger joint angle with a root-mean-square error of 29-34 degrees. A multiple regression model with a second-order polynomial basis function was shown to be suitable for the estimation of all total finger joint angles, but not those of the thumb.

Methods for estimating tributary streamflow in the Chattahoochee River basin between Buford Dam and Franklin, Georgia

USGS Publications Warehouse

Stamey, Timothy C.

1998-01-01

Simple and reliable methods for estimating hourly streamflow are needed for the calibration and verification of a Chattahoochee River basin model between Buford Dam and Franklin, Ga. The river basin model is being developed by Georgia Department of Natural Resources, Environmental Protection Division, as part of their Chattahoochee River Modeling Project. Concurrent streamflow data collected at 19 continuous-record, and 31 partial-record streamflow stations, were used in ordinary least-squares linear regression analyses to define estimating equations, and in verifying drainage-area prorations. The resulting regression or drainage-area ratio estimating equations were used to compute hourly streamflow at the partial-record stations. The coefficients of determination (r-squared values) for the regression estimating equations ranged from 0.90 to 0.99. Observed and estimated hourly and daily streamflow data were computed for May 1, 1995, through October 31, 1995. Comparisons of observed and estimated daily streamflow data for 12 continuous-record tributary stations, that had available streamflow data for all or part of the period from May 1, 1995, to October 31, 1995, indicate that the mean error of estimate for the daily streamflow was about 25 percent.

Comparison of partial least squares and random forests for evaluating relationship between phenolics and bioactivities of Neptunia oleracea.

PubMed

Lee, Soo Yee; Mediani, Ahmed; Maulidiani, Maulidiani; Khatib, Alfi; Ismail, Intan Safinar; Zawawi, Norhasnida; Abas, Faridah

2018-01-01

Neptunia oleracea is a plant consumed as a vegetable and which has been used as a folk remedy for several diseases. Herein, two regression models (partial least squares, PLS; and random forest, RF) in a metabolomics approach were compared and applied to the evaluation of the relationship between phenolics and bioactivities of N. oleracea. In addition, the effects of different extraction conditions on the phenolic constituents were assessed by pattern recognition analysis. Comparison of the PLS and RF showed that RF exhibited poorer generalization and hence poorer predictive performance. Both the regression coefficient of PLS and the variable importance of RF revealed that quercetin and kaempferol derivatives, caffeic acid and vitexin-2-O-rhamnoside were significant towards the tested bioactivities. Furthermore, principal component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA) results showed that sonication and absolute ethanol are the preferable extraction method and ethanol ratio, respectively, to produce N. oleracea extracts with high phenolic levels and therefore high DPPH scavenging and α-glucosidase inhibitory activities. Both PLS and RF are useful regression models in metabolomics studies. This work provides insight into the performance of different multivariate data analysis tools and the effects of different extraction conditions on the extraction of desired phenolics from plants. © 2017 Society of Chemical Industry. © 2017 Society of Chemical Industry.

Comparative Analysis of River Flow Modelling by Using Supervised Learning Technique

NASA Astrophysics Data System (ADS)

Ismail, Shuhaida; Mohamad Pandiahi, Siraj; Shabri, Ani; Mustapha, Aida

2018-04-01

The goal of this research is to investigate the efficiency of three supervised learning algorithms for forecasting monthly river flow of the Indus River in Pakistan, spread over 550 square miles or 1800 square kilometres. The algorithms include the Least Square Support Vector Machine (LSSVM), Artificial Neural Network (ANN) and Wavelet Regression (WR). The forecasting models predict the monthly river flow obtained from the three models individually for river flow data and the accuracy of the all models were then compared against each other. The monthly river flow of the said river has been forecasted using these three models. The obtained results were compared and statistically analysed. Then, the results of this analytical comparison showed that LSSVM model is more precise in the monthly river flow forecasting. It was found that LSSVM has he higher r with the value of 0.934 compared to other models. This indicate that LSSVM is more accurate and efficient as compared to the ANN and WR model.

Nondestructive evaluation of soluble solid content in strawberry by near infrared spectroscopy

NASA Astrophysics Data System (ADS)

Guo, Zhiming; Huang, Wenqian; Chen, Liping; Wang, Xiu; Peng, Yankun

This paper indicates the feasibility to use near infrared (NIR) spectroscopy combined with synergy interval partial least squares (siPLS) algorithms as a rapid nondestructive method to estimate the soluble solid content (SSC) in strawberry. Spectral preprocessing methods were optimized selected by cross-validation in the model calibration. Partial least squares (PLS) algorithm was conducted on the calibration of regression model. The performance of the final model was back-evaluated according to root mean square error of calibration (RMSEC) and correlation coefficient (R2 c) in calibration set, and tested by mean square error of prediction (RMSEP) and correlation coefficient (R2 p) in prediction set. The optimal siPLS model was obtained with after first derivation spectra preprocessing. The measurement results of best model were achieved as follow: RMSEC = 0.2259, R2 c = 0.9590 in the calibration set; and RMSEP = 0.2892, R2 p = 0.9390 in the prediction set. This work demonstrated that NIR spectroscopy and siPLS with efficient spectral preprocessing is a useful tool for nondestructively evaluation SSC in strawberry.

Use of partial least squares regression to impute SNP genotypes in Italian cattle breeds.

PubMed

Dimauro, Corrado; Cellesi, Massimo; Gaspa, Giustino; Ajmone-Marsan, Paolo; Steri, Roberto; Marras, Gabriele; Macciotta, Nicolò P P

2013-06-05

The objective of the present study was to test the ability of the partial least squares regression technique to impute genotypes from low density single nucleotide polymorphisms (SNP) panels i.e. 3K or 7K to a high density panel with 50K SNP. No pedigree information was used. Data consisted of 2093 Holstein, 749 Brown Swiss and 479 Simmental bulls genotyped with the Illumina 50K Beadchip. First, a single-breed approach was applied by using only data from Holstein animals. Then, to enlarge the training population, data from the three breeds were combined and a multi-breed analysis was performed. Accuracies of genotypes imputed using the partial least squares regression method were compared with those obtained by using the Beagle software. The impact of genotype imputation on breeding value prediction was evaluated for milk yield, fat content and protein content. In the single-breed approach, the accuracy of imputation using partial least squares regression was around 90 and 94% for the 3K and 7K platforms, respectively; corresponding accuracies obtained with Beagle were around 85% and 90%. Moreover, computing time required by the partial least squares regression method was on average around 10 times lower than computing time required by Beagle. Using the partial least squares regression method in the multi-breed resulted in lower imputation accuracies than using single-breed data. The impact of the SNP-genotype imputation on the accuracy of direct genomic breeding values was small. The correlation between estimates of genetic merit obtained by using imputed versus actual genotypes was around 0.96 for the 7K chip. Results of the present work suggested that the partial least squares regression imputation method could be useful to impute SNP genotypes when pedigree information is not available.

Assessing the sensitivity and robustness of prediction models for apple firmness using spectral scattering technique

USDA-ARS?s Scientific Manuscript database

Spectral scattering is useful for nondestructive sensing of fruit firmness. Prediction models, however, are typically built using multivariate statistical methods such as partial least squares regression (PLSR), whose performance generally depends on the characteristics of the data. The aim of this ...

Web-based Interspecies Correlation Estimation (Web-ICE) for Acute Toxicity: User Manual Version 3.1

EPA Science Inventory

Predictive toxicological models are integral to ecological risk assessment because data for most species are limited. Web-based Interspecies Correlation Estimation (Web-ICE) models are least square regressions that predict acute toxicity (LC50/LD50) of a chemical to a species, ge...

WEB-BASED INTERSPECIES CORRELATION ESTIMATION (WEB-ICE) FOR ACUTE TOXICITY: USER MANUAL V2

EPA Science Inventory

Predictive toxicological models are integral to environmental risk Assessment where data for most species are limited. Web-based Interspecies Correlation Estimation (Web-ICE) models are least square regressions that predict acute toxicity (LC50/LD50) of a chemical to a species, ...

Individual tree growth models for natural even-aged shortleaf pine

Treesearch

Chakra B. Budhathoki; Thomas B. Lynch; James M. Guldin

2006-01-01

Shortleaf pine (Pinus echinata Mill.) measurements were available from permanent plots established in even-aged stands of the Ouachita Mountains for studying growth. Annual basal area growth was modeled with a least-squares nonlinear regression method utilizing three measurements. The analysis showed that the parameter estimates were in agreement...

MERGANSER - An Empirical Model to Predict Fish and Loon Mercury in New England Lakes

EPA Science Inventory

MERGANSER (MERcury Geo-spatial AssessmeNtS for the New England Region) is an empirical least-squares multiple regression model using mercury (Hg) deposition and readily obtainable lake and watershed features to predict fish (fillet) and common loon (blood) Hg in New England lakes...

Comparison of Peak-Flow Estimation Methods for Small Drainage Basins in Maine

USGS Publications Warehouse

Hodgkins, Glenn A.; Hebson, Charles; Lombard, Pamela J.; Mann, Alexander

2007-01-01

Understanding the accuracy of commonly used methods for estimating peak streamflows is important because the designs of bridges, culverts, and other river structures are based on these flows. Different methods for estimating peak streamflows were analyzed for small drainage basins in Maine. For the smallest basins, with drainage areas of 0.2 to 1.0 square mile, nine peak streamflows from actual rainfall events at four crest-stage gaging stations were modeled by the Rational Method and the Natural Resource Conservation Service TR-20 method and compared to observed peak flows. The Rational Method had a root mean square error (RMSE) of -69.7 to 230 percent (which means that approximately two thirds of the modeled flows were within -69.7 to 230 percent of the observed flows). The TR-20 method had an RMSE of -98.0 to 5,010 percent. Both the Rational Method and TR-20 underestimated the observed flows in most cases. For small basins, with drainage areas of 1.0 to 10 square miles, modeled peak flows were compared to observed statistical peak flows with return periods of 2, 50, and 100 years for 17 streams in Maine and adjoining parts of New Hampshire. Peak flows were modeled by the Rational Method, the Natural Resources Conservation Service TR-20 method, U.S. Geological Survey regression equations, and the Probabilistic Rational Method. The regression equations were the most accurate method of computing peak flows in Maine for streams with drainage areas of 1.0 to 10 square miles with an RMSE of -34.3 to 52.2 percent for 50-year peak flows. The Probabilistic Rational Method was the next most accurate method (-38.5 to 62.6 percent). The Rational Method (-56.1 to 128 percent) and particularly the TR-20 method (-76.4 to 323 percent) had much larger errors. Both the TR-20 and regression methods had similar numbers of underpredictions and overpredictions. The Rational Method overpredicted most peak flows and the Probabilistic Rational Method tended to overpredict peak flows from the smaller (less than 5 square miles) drainage basins and underpredict peak flows from larger drainage basins. The results of this study are consistent with the most comprehensive analysis of observed and modeled peak streamflows in the United States, which analyzed statistical peak flows from 70 drainage basins in the Midwest and the Northwest.

Environmental fate model for ultra-low-volume insecticide applications used for adult mosquito management

USGS Publications Warehouse

Schleier, Jerome J.; Peterson, Robert K.D.; Irvine, Kathryn M.; Marshall, Lucy M.; Weaver, David K.; Preftakes, Collin J.

2012-01-01

One of the more effective ways of managing high densities of adult mosquitoes that vector human and animal pathogens is ultra-low-volume (ULV) aerosol applications of insecticides. The U.S. Environmental Protection Agency uses models that are not validated for ULV insecticide applications and exposure assumptions to perform their human and ecological risk assessments. Currently, there is no validated model that can accurately predict deposition of insecticides applied using ULV technology for adult mosquito management. In addition, little is known about the deposition and drift of small droplets like those used under conditions encountered during ULV applications. The objective of this study was to perform field studies to measure environmental concentrations of insecticides and to develop a validated model to predict the deposition of ULV insecticides. The final regression model was selected by minimizing the Bayesian Information Criterion and its prediction performance was evaluated using k-fold cross validation. Density of the formulation and the density and CMD interaction coefficients were the largest in the model. The results showed that as density of the formulation decreases, deposition increases. The interaction of density and CMD showed that higher density formulations and larger droplets resulted in greater deposition. These results are supported by the aerosol physics literature. A k-fold cross validation demonstrated that the mean square error of the selected regression model is not biased, and the mean square error and mean square prediction error indicated good predictive ability.

Examination of Parameters Affecting the House Prices by Multiple Regression Analysis and its Contributions to Earthquake-Based Urban Transformation

NASA Astrophysics Data System (ADS)

Denli, H. H.; Durmus, B.

2016-12-01

The purpose of this study is to examine the factors which may affect the apartment prices with multiple linear regression analysis models and visualize the results by value maps. The study is focused on a county of Istanbul - Turkey. Totally 390 apartments around the county Umraniye are evaluated due to their physical and locational conditions. The identification of factors affecting the price of apartments in the county with a population of approximately 600k is expected to provide a significant contribution to the apartment market.Physical factors are selected as the age, number of rooms, size, floor numbers of the building and the floor that the apartment is positioned in. Positional factors are selected as the distances to the nearest hospital, school, park and police station. Totally ten physical and locational parameters are examined by regression analysis.After the regression analysis has been performed, value maps are composed from the parameters age, price and price per square meters. The most significant of the composed maps is the price per square meters map. Results show that the location of the apartment has the most influence to the square meter price information of the apartment. A different practice is developed from the composed maps by searching the ability of using price per square meters map in urban transformation practices. By marking the buildings older than 15 years in the price per square meters map, a different and new interpretation has been made to determine the buildings, to which should be given priority during an urban transformation in the county.This county is very close to the North Anatolian Fault zone and is under the threat of earthquakes. By marking the apartments older than 15 years on the price per square meters map, both older and expensive square meters apartments list can be gathered. By the help of this list, the priority could be given to the selected higher valued old apartments to support the economy of the country during an earthquake loss. We may call this urban transformation as earthquake-based urban transformation.

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.

Detection of melamine in milk powders using Near-Infrared Hyperspectral imaging combined with regression coefficient of partial least square regression model

USDA-ARS?s Scientific Manuscript database

Illegal use of nitrogen-rich melamine (C3H6N6) to boost perceived protein content of food products such as milk, infant formula, frozen yogurt, pet food, biscuits, and coffee drinks has caused serious food safety problems. Conventional methods to detect melamine in foods, such as Enzyme-linked immun...

An Application of Interactive Computer Graphics to the Study of Inferential Statistics and the General Linear Model

DTIC Science & Technology

1991-09-01

matrix, the Regression Sum of Squares (SSR) and Error Sum of Squares (SSE) are also displayed as a percentage of the Total Sum of Squares ( SSTO ...vector when the student compares the SSR to the SSE. In addition to the plot, the actual values of SSR, SSE, and SSTO are also provided. Figure 3 gives the...Es ainSpace = E 3 Error- Eor Space =n t! L . Pro~cio q Yonto Pro~rct on of Y onto the simaton, pac ror Space SSR SSEL0.20 IV = 14,1 +IErrorI 2 SSTO

Cost-sensitive AdaBoost algorithm for ordinal regression based on extreme learning machine.

PubMed

Riccardi, Annalisa; Fernández-Navarro, Francisco; Carloni, Sante

2014-10-01

In this paper, the well known stagewise additive modeling using a multiclass exponential (SAMME) boosting algorithm is extended to address problems where there exists a natural order in the targets using a cost-sensitive approach. The proposed ensemble model uses an extreme learning machine (ELM) model as a base classifier (with the Gaussian kernel and the additional regularization parameter). The closed form of the derived weighted least squares problem is provided, and it is employed to estimate analytically the parameters connecting the hidden layer to the output layer at each iteration of the boosting algorithm. Compared to the state-of-the-art boosting algorithms, in particular those using ELM as base classifier, the suggested technique does not require the generation of a new training dataset at each iteration. The adoption of the weighted least squares formulation of the problem has been presented as an unbiased and alternative approach to the already existing ELM boosting techniques. Moreover, the addition of a cost model for weighting the patterns, according to the order of the targets, enables the classifier to tackle ordinal regression problems further. The proposed method has been validated by an experimental study by comparing it with already existing ensemble methods and ELM techniques for ordinal regression, showing competitive results.

High and low frequency unfolded partial least squares regression based on empirical mode decomposition for quantitative analysis of fuel oil samples.

PubMed

Bian, Xihui; Li, Shujuan; Lin, Ligang; Tan, Xiaoyao; Fan, Qingjie; Li, Ming

2016-06-21

Accurate prediction of the model is fundamental to the successful analysis of complex samples. To utilize abundant information embedded over frequency and time domains, a novel regression model is presented for quantitative analysis of hydrocarbon contents in the fuel oil samples. The proposed method named as high and low frequency unfolded PLSR (HLUPLSR), which integrates empirical mode decomposition (EMD) and unfolded strategy with partial least squares regression (PLSR). In the proposed method, the original signals are firstly decomposed into a finite number of intrinsic mode functions (IMFs) and a residue by EMD. Secondly, the former high frequency IMFs are summed as a high frequency matrix and the latter IMFs and residue are summed as a low frequency matrix. Finally, the two matrices are unfolded to an extended matrix in variable dimension, and then the PLSR model is built between the extended matrix and the target values. Coupled with Ultraviolet (UV) spectroscopy, HLUPLSR has been applied to determine hydrocarbon contents of light gas oil and diesel fuels samples. Comparing with single PLSR and other signal processing techniques, the proposed method shows superiority in prediction ability and better model interpretation. Therefore, HLUPLSR method provides a promising tool for quantitative analysis of complex samples. Copyright © 2016 Elsevier B.V. All rights reserved.

Hypothesis Testing Using Factor Score Regression

PubMed Central

Devlieger, Ines; Mayer, Axel; Rosseel, Yves

2015-01-01

In this article, an overview is given of four methods to perform factor score regression (FSR), namely regression FSR, Bartlett FSR, the bias avoiding method of Skrondal and Laake, and the bias correcting method of Croon. The bias correcting method is extended to include a reliable standard error. The four methods are compared with each other and with structural equation modeling (SEM) by using analytic calculations and two Monte Carlo simulation studies to examine their finite sample characteristics. Several performance criteria are used, such as the bias using the unstandardized and standardized parameterization, efficiency, mean square error, standard error bias, type I error rate, and power. The results show that the bias correcting method, with the newly developed standard error, is the only suitable alternative for SEM. While it has a higher standard error bias than SEM, it has a comparable bias, efficiency, mean square error, power, and type I error rate. PMID:29795886

Measurement System Characterization in the Presence of Measurement Errors

NASA Technical Reports Server (NTRS)

Commo, Sean A.

2012-01-01

In the calibration of a measurement system, data are collected in order to estimate a mathematical model between one or more factors of interest and a response. Ordinary least squares is a method employed to estimate the regression coefficients in the model. The method assumes that the factors are known without error; yet, it is implicitly known that the factors contain some uncertainty. In the literature, this uncertainty is known as measurement error. The measurement error affects both the estimates of the model coefficients and the prediction, or residual, errors. There are some methods, such as orthogonal least squares, that are employed in situations where measurement errors exist, but these methods do not directly incorporate the magnitude of the measurement errors. This research proposes a new method, known as modified least squares, that combines the principles of least squares with knowledge about the measurement errors. This knowledge is expressed in terms of the variance ratio - the ratio of response error variance to measurement error variance.

Estimating EQ-5D values from the Neck Disability Index and numeric rating scales for neck and arm pain.

PubMed

Carreon, Leah Y; Bratcher, Kelly R; Das, Nandita; Nienhuis, Jacob B; Glassman, Steven D

2014-09-01

The Neck Disability Index (NDI) and numeric rating scales (0 to 10) for neck pain and arm pain are widely used cervical spine disease-specific measures. Recent studies have shown that there is a strong relationship between the SF-6D and the NDI such that using a simple linear regression allows for the estimation of an SF-6D value from the NDI alone. Due to ease of administration and scoring, the EQ-5D is increasingly being used as a measure of utility in the clinical setting. The purpose of this study is to determine if the EQ-5D values can be estimated from commonly available cervical spine disease-specific health-related quality of life measures, much like the SF-6D. The EQ-5D, NDI, neck pain score, and arm pain score were prospectively collected in 3732 patients who presented to the authors' clinic with degenerative cervical spine disorders. Correlation coefficients for paired observations from multiple time points between the NDI, neck pain and arm pain scores, and EQ-5D were determined. Regression models were built to estimate the EQ-5D values from the NDI, neck pain, and arm pain scores. The mean age of the 3732 patients was 53.3 ± 12.2 years, and 43% were male. Correlations between the EQ-5D and the NDI, neck pain score, and arm pain score were statistically significant (p < 0.0001), with correlation coefficients of -0.77, -0.62, and -0.50, respectively. The regression equation 0.98947 + (-0.00705 × NDI) + (-0.00875 × arm pain score) + (-0.00877 × neck pain score) to predict EQ-5D had an R-square of 0.62 and a root mean square error (RMSE) of 0.146. The model using NDI alone had an R-square of 0.59 and a RMSE of 0.150. The model using the individual NDI items had an R-square of 0.46 and an RMSE of 0.172. The correlation coefficient between the observed and estimated EQ-5D scores was 0.79. There was no statistically significant difference between the actual EQ-5D score (0.603 ± 0.235) and the estimated EQ-5D score (0.603 ± 0.185) using the NDI, neck pain score, and arm pain score regression model. However, rounding off the coefficients to fewer than 5 decimal places produced less accurate results. The regression model estimating the EQ-5D from the NDI, neck pain score, and arm pain score accounted for 60% of the variability of the EQ-5D with a relatively large RMSE. This regression model may not be sufficient to accurately or reliably estimate actual EQ-5D values.

Experimental and computational prediction of glass transition temperature of drugs.

PubMed

Alzghoul, Ahmad; Alhalaweh, Amjad; Mahlin, Denny; Bergström, Christel A S

2014-12-22

Glass transition temperature (Tg) is an important inherent property of an amorphous solid material which is usually determined experimentally. In this study, the relation between Tg and melting temperature (Tm) was evaluated using a data set of 71 structurally diverse druglike compounds. Further, in silico models for prediction of Tg were developed based on calculated molecular descriptors and linear (multilinear regression, partial least-squares, principal component regression) and nonlinear (neural network, support vector regression) modeling techniques. The models based on Tm predicted Tg with an RMSE of 19.5 K for the test set. Among the five computational models developed herein the support vector regression gave the best result with RMSE of 18.7 K for the test set using only four chemical descriptors. Hence, two different models that predict Tg of drug-like molecules with high accuracy were developed. If Tm is available, a simple linear regression can be used to predict Tg. However, the results also suggest that support vector regression and calculated molecular descriptors can predict Tg with equal accuracy, already before compound synthesis.

The prediction of food additives in the fruit juice based on electronic nose with chemometrics.

PubMed

Qiu, Shanshan; Wang, Jun

2017-09-01

Food additives are added to products to enhance their taste, and preserve flavor or appearance. While their use should be restricted to achieve a technological benefit, the contents of food additives should be also strictly controlled. In this study, E-nose was applied as an alternative to traditional monitoring technologies for determining two food additives, namely benzoic acid and chitosan. For quantitative monitoring, support vector machine (SVM), random forest (RF), extreme learning machine (ELM) and partial least squares regression (PLSR) were applied to establish regression models between E-nose signals and the amount of food additives in fruit juices. The monitoring models based on ELM and RF reached higher correlation coefficients (R 2 s) and lower root mean square errors (RMSEs) than models based on PLSR and SVM. This work indicates that E-nose combined with RF or ELM can be a cost-effective, easy-to-build and rapid detection system for food additive monitoring. Copyright © 2017 Elsevier Ltd. All rights reserved.

Sugar and acid content of Citrus prediction modeling using FT-IR fingerprinting in combination with multivariate statistical analysis.

PubMed

Song, Seung Yeob; Lee, Young Koung; Kim, In-Jung

2016-01-01

A high-throughput screening system for Citrus lines were established with higher sugar and acid contents using Fourier transform infrared (FT-IR) spectroscopy in combination with multivariate analysis. FT-IR spectra confirmed typical spectral differences between the frequency regions of 950-1100 cm(-1), 1300-1500 cm(-1), and 1500-1700 cm(-1). Principal component analysis (PCA) and subsequent partial least square-discriminant analysis (PLS-DA) were able to discriminate five Citrus lines into three separate clusters corresponding to their taxonomic relationships. The quantitative predictive modeling of sugar and acid contents from Citrus fruits was established using partial least square regression algorithms from FT-IR spectra. The regression coefficients (R(2)) between predicted values and estimated sugar and acid content values were 0.99. These results demonstrate that by using FT-IR spectra and applying quantitative prediction modeling to Citrus sugar and acid contents, excellent Citrus lines can be early detected with greater accuracy. Copyright © 2015 Elsevier Ltd. All rights reserved.

Analyzing industrial energy use through ordinary least squares regression models

NASA Astrophysics Data System (ADS)

Golden, Allyson Katherine

Extensive research has been performed using regression analysis and calibrated simulations to create baseline energy consumption models for residential buildings and commercial institutions. However, few attempts have been made to discuss the applicability of these methodologies to establish baseline energy consumption models for industrial manufacturing facilities. In the few studies of industrial facilities, the presented linear change-point and degree-day regression analyses illustrate ideal cases. It follows that there is a need in the established literature to discuss the methodologies and to determine their applicability for establishing baseline energy consumption models of industrial manufacturing facilities. The thesis determines the effectiveness of simple inverse linear statistical regression models when establishing baseline energy consumption models for industrial manufacturing facilities. Ordinary least squares change-point and degree-day regression methods are used to create baseline energy consumption models for nine different case studies of industrial manufacturing facilities located in the southeastern United States. The influence of ambient dry-bulb temperature and production on total facility energy consumption is observed. The energy consumption behavior of industrial manufacturing facilities is only sometimes sufficiently explained by temperature, production, or a combination of the two variables. This thesis also provides methods for generating baseline energy models that are straightforward and accessible to anyone in the industrial manufacturing community. The methods outlined in this thesis may be easily replicated by anyone that possesses basic spreadsheet software and general knowledge of the relationship between energy consumption and weather, production, or other influential variables. With the help of simple inverse linear regression models, industrial manufacturing facilities may better understand their energy consumption and production behavior, and identify opportunities for energy and cost savings. This thesis study also utilizes change-point and degree-day baseline energy models to disaggregate facility annual energy consumption into separate industrial end-user categories. The baseline energy model provides a suitable and economical alternative to sub-metering individual manufacturing equipment. One case study describes the conjoined use of baseline energy models and facility information gathered during a one-day onsite visit to perform an end-point energy analysis of an injection molding facility conducted by the Alabama Industrial Assessment Center. Applying baseline regression model results to the end-point energy analysis allowed the AIAC to better approximate the annual energy consumption of the facility's HVAC system.

Rapid detection of talcum powder in tea using FT-IR spectroscopy coupled with chemometrics

PubMed Central

Li, Xiaoli; Zhang, Yuying; He, Yong

2016-01-01

This paper investigated the feasibility of Fourier transform infrared transmission (FT-IR) spectroscopy to detect talcum powder illegally added in tea based on chemometric methods. Firstly, 210 samples of tea powder with 13 dose levels of talcum powder were prepared for FT-IR spectra acquirement. In order to highlight the slight variations in FT-IR spectra, smoothing, normalize and standard normal variate (SNV) were employed to preprocess the raw spectra. Among them, SNV preprocessing had the best performance with high correlation of prediction (RP = 0.948) and low root mean square error of prediction (RMSEP = 0.108) of partial least squares (PLS) model. Then 18 characteristic wavenumbers were selected based on a hybrid of backward interval partial least squares (biPLS) regression, competitive adaptive reweighted sampling (CARS) algorithm and successive projections algorithm (SPA). These characteristic wavenumbers only accounted for 0.64% of the full wavenumbers. Following that, 18 characteristic wavenumbers were used to build linear and nonlinear determination models by PLS regression and extreme learning machine (ELM), respectively. The optimal model with RP = 0.963 and RMSEP = 0.137 was achieved by ELM algorithm. These results demonstrated that FT-IR spectroscopy with chemometrics could be used successfully to detect talcum powder in tea. PMID:27468701

Prediction models for CO2 emission in Malaysia using best subsets regression and multi-linear regression

NASA Astrophysics Data System (ADS)

Tan, C. H.; Matjafri, M. Z.; Lim, H. S.

2015-10-01

This paper presents the prediction models which analyze and compute the CO2 emission in Malaysia. Each prediction model for CO2 emission will be analyzed based on three main groups which is transportation, electricity and heat production as well as residential buildings and commercial and public services. The prediction models were generated using data obtained from World Bank Open Data. Best subset method will be used to remove irrelevant data and followed by multi linear regression to produce the prediction models. From the results, high R-square (prediction) value was obtained and this implies that the models are reliable to predict the CO2 emission by using specific data. In addition, the CO2 emissions from these three groups are forecasted using trend analysis plots for observation purpose.

Local polynomial estimation of heteroscedasticity in a multivariate linear regression model and its applications in economics.

PubMed

Su, Liyun; Zhao, Yanyong; Yan, Tianshun; Li, Fenglan

2012-01-01

Multivariate local polynomial fitting is applied to the multivariate linear heteroscedastic regression model. Firstly, the local polynomial fitting is applied to estimate heteroscedastic function, then the coefficients of regression model are obtained by using generalized least squares method. One noteworthy feature of our approach is that we avoid the testing for heteroscedasticity by improving the traditional two-stage method. Due to non-parametric technique of local polynomial estimation, it is unnecessary to know the form of heteroscedastic function. Therefore, we can improve the estimation precision, when the heteroscedastic function is unknown. Furthermore, we verify that the regression coefficients is asymptotic normal based on numerical simulations and normal Q-Q plots of residuals. Finally, the simulation results and the local polynomial estimation of real data indicate that our approach is surely effective in finite-sample situations.

Improvement of near infrared spectroscopic (NIRS) analysis of caffeine in roasted Arabica coffee by variable selection method of stability competitive adaptive reweighted sampling (SCARS).

PubMed

Zhang, Xuan; Li, Wei; Yin, Bin; Chen, Weizhong; Kelly, Declan P; Wang, Xiaoxin; Zheng, Kaiyi; Du, Yiping

2013-10-01

Coffee is the most heavily consumed beverage in the world after water, for which quality is a key consideration in commercial trade. Therefore, caffeine content which has a significant effect on the final quality of the coffee products requires to be determined fast and reliably by new analytical techniques. The main purpose of this work was to establish a powerful and practical analytical method based on near infrared spectroscopy (NIRS) and chemometrics for quantitative determination of caffeine content in roasted Arabica coffees. Ground coffee samples within a wide range of roasted levels were analyzed by NIR, meanwhile, in which the caffeine contents were quantitative determined by the most commonly used HPLC-UV method as the reference values. Then calibration models based on chemometric analyses of the NIR spectral data and reference concentrations of coffee samples were developed. Partial least squares (PLS) regression was used to construct the models. Furthermore, diverse spectra pretreatment and variable selection techniques were applied in order to obtain robust and reliable reduced-spectrum regression models. Comparing the respective quality of the different models constructed, the application of second derivative pretreatment and stability competitive adaptive reweighted sampling (SCARS) variable selection provided a notably improved regression model, with root mean square error of cross validation (RMSECV) of 0.375 mg/g and correlation coefficient (R) of 0.918 at PLS factor of 7. An independent test set was used to assess the model, with the root mean square error of prediction (RMSEP) of 0.378 mg/g, mean relative error of 1.976% and mean relative standard deviation (RSD) of 1.707%. Thus, the results provided by the high-quality calibration model revealed the feasibility of NIR spectroscopy for at-line application to predict the caffeine content of unknown roasted coffee samples, thanks to the short analysis time of a few seconds and non-destructive advantages of NIRS. Copyright © 2013 Elsevier B.V. All rights reserved.

Improvement of near infrared spectroscopic (NIRS) analysis of caffeine in roasted Arabica coffee by variable selection method of stability competitive adaptive reweighted sampling (SCARS)

NASA Astrophysics Data System (ADS)

Zhang, Xuan; Li, Wei; Yin, Bin; Chen, Weizhong; Kelly, Declan P.; Wang, Xiaoxin; Zheng, Kaiyi; Du, Yiping

2013-10-01

Coffee is the most heavily consumed beverage in the world after water, for which quality is a key consideration in commercial trade. Therefore, caffeine content which has a significant effect on the final quality of the coffee products requires to be determined fast and reliably by new analytical techniques. The main purpose of this work was to establish a powerful and practical analytical method based on near infrared spectroscopy (NIRS) and chemometrics for quantitative determination of caffeine content in roasted Arabica coffees. Ground coffee samples within a wide range of roasted levels were analyzed by NIR, meanwhile, in which the caffeine contents were quantitative determined by the most commonly used HPLC-UV method as the reference values. Then calibration models based on chemometric analyses of the NIR spectral data and reference concentrations of coffee samples were developed. Partial least squares (PLS) regression was used to construct the models. Furthermore, diverse spectra pretreatment and variable selection techniques were applied in order to obtain robust and reliable reduced-spectrum regression models. Comparing the respective quality of the different models constructed, the application of second derivative pretreatment and stability competitive adaptive reweighted sampling (SCARS) variable selection provided a notably improved regression model, with root mean square error of cross validation (RMSECV) of 0.375 mg/g and correlation coefficient (R) of 0.918 at PLS factor of 7. An independent test set was used to assess the model, with the root mean square error of prediction (RMSEP) of 0.378 mg/g, mean relative error of 1.976% and mean relative standard deviation (RSD) of 1.707%. Thus, the results provided by the high-quality calibration model revealed the feasibility of NIR spectroscopy for at-line application to predict the caffeine content of unknown roasted coffee samples, thanks to the short analysis time of a few seconds and non-destructive advantages of NIRS.

Generalized regression neural network (GRNN)-based approach for colored dissolved organic matter (CDOM) retrieval: case study of Connecticut River at Middle Haddam Station, USA.

PubMed

Heddam, Salim

2014-11-01

The prediction of colored dissolved organic matter (CDOM) using artificial neural network approaches has received little attention in the past few decades. In this study, colored dissolved organic matter (CDOM) was modeled using generalized regression neural network (GRNN) and multiple linear regression (MLR) models as a function of Water temperature (TE), pH, specific conductance (SC), and turbidity (TU). Evaluation of the prediction accuracy of the models is based on the root mean square error (RMSE), mean absolute error (MAE), coefficient of correlation (CC), and Willmott's index of agreement (d). The results indicated that GRNN can be applied successfully for prediction of colored dissolved organic matter (CDOM).

Least squares regression methods for clustered ROC data with discrete covariates.

PubMed

Tang, Liansheng Larry; Zhang, Wei; Li, Qizhai; Ye, Xuan; Chan, Leighton

2016-07-01

The receiver operating characteristic (ROC) curve is a popular tool to evaluate and compare the accuracy of diagnostic tests to distinguish the diseased group from the nondiseased group when test results from tests are continuous or ordinal. A complicated data setting occurs when multiple tests are measured on abnormal and normal locations from the same subject and the measurements are clustered within the subject. Although least squares regression methods can be used for the estimation of ROC curve from correlated data, how to develop the least squares methods to estimate the ROC curve from the clustered data has not been studied. Also, the statistical properties of the least squares methods under the clustering setting are unknown. In this article, we develop the least squares ROC methods to allow the baseline and link functions to differ, and more importantly, to accommodate clustered data with discrete covariates. The methods can generate smooth ROC curves that satisfy the inherent continuous property of the true underlying curve. The least squares methods are shown to be more efficient than the existing nonparametric ROC methods under appropriate model assumptions in simulation studies. We apply the methods to a real example in the detection of glaucomatous deterioration. We also derive the asymptotic properties of the proposed methods. © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

Detection of melamine in milk powders using near-infrared hyperspectral imaging combined with regression coefficient of partial least square regression model.

PubMed

Lim, Jongguk; Kim, Giyoung; Mo, Changyeun; Kim, Moon S; Chao, Kuanglin; Qin, Jianwei; Fu, Xiaping; Baek, Insuck; Cho, Byoung-Kwan

2016-05-01

Illegal use of nitrogen-rich melamine (C3H6N6) to boost perceived protein content of food products such as milk, infant formula, frozen yogurt, pet food, biscuits, and coffee drinks has caused serious food safety problems. Conventional methods to detect melamine in foods, such as Enzyme-linked immunosorbent assay (ELISA), High-performance liquid chromatography (HPLC), and Gas chromatography-mass spectrometry (GC-MS), are sensitive but they are time-consuming, expensive, and labor-intensive. In this research, near-infrared (NIR) hyperspectral imaging technique combined with regression coefficient of partial least squares regression (PLSR) model was used to detect melamine particles in milk powders easily and quickly. NIR hyperspectral reflectance imaging data in the spectral range of 990-1700nm were acquired from melamine-milk powder mixture samples prepared at various concentrations ranging from 0.02% to 1%. PLSR models were developed to correlate the spectral data (independent variables) with melamine concentration (dependent variables) in melamine-milk powder mixture samples. PLSR models applying various pretreatment methods were used to reconstruct the two-dimensional PLS images. PLS images were converted to the binary images to detect the suspected melamine pixels in milk powder. As the melamine concentration was increased, the numbers of suspected melamine pixels of binary images were also increased. These results suggested that NIR hyperspectral imaging technique and the PLSR model can be regarded as an effective tool to detect melamine particles in milk powders. Copyright © 2016 Elsevier B.V. All rights reserved.

A comparison of model-based imputation methods for handling missing predictor values in a linear regression model: A simulation study

NASA Astrophysics Data System (ADS)

Hasan, Haliza; Ahmad, Sanizah; Osman, Balkish Mohd; Sapri, Shamsiah; Othman, Nadirah

2017-08-01

In regression analysis, missing covariate data has been a common problem. Many researchers use ad hoc methods to overcome this problem due to the ease of implementation. However, these methods require assumptions about the data that rarely hold in practice. Model-based methods such as Maximum Likelihood (ML) using the expectation maximization (EM) algorithm and Multiple Imputation (MI) are more promising when dealing with difficulties caused by missing data. Then again, inappropriate methods of missing value imputation can lead to serious bias that severely affects the parameter estimates. The main objective of this study is to provide a better understanding regarding missing data concept that can assist the researcher to select the appropriate missing data imputation methods. A simulation study was performed to assess the effects of different missing data techniques on the performance of a regression model. The covariate data were generated using an underlying multivariate normal distribution and the dependent variable was generated as a combination of explanatory variables. Missing values in covariate were simulated using a mechanism called missing at random (MAR). Four levels of missingness (10%, 20%, 30% and 40%) were imposed. ML and MI techniques available within SAS software were investigated. A linear regression analysis was fitted and the model performance measures; MSE, and R-Squared were obtained. Results of the analysis showed that MI is superior in handling missing data with highest R-Squared and lowest MSE when percent of missingness is less than 30%. Both methods are unable to handle larger than 30% level of missingness.

Normality of Residuals Is a Continuous Variable, and Does Seem to Influence the Trustworthiness of Confidence Intervals: A Response to, and Appreciation of, Williams, Grajales, and Kurkiewicz (2013)

ERIC Educational Resources Information Center

Osborne, Jason W.

2013-01-01

Osborne and Waters (2002) focused on checking some of the assumptions of multiple linear regression. In a critique of that paper, Williams, Grajales, and Kurkiewicz correctly clarify that regression models estimated using ordinary least squares require the assumption of normally distributed errors, but not the assumption of normally distributed…

Analytics of Radioactive Materials Released in the Fukushima Daiichi Nuclear Accident

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

Egarievwe, Stephen U.; Nuclear Engineering Department, University of Tennessee, Knoxville, TN; Coble, Jamie B.

The 2011 Fukushima Daiichi nuclear accident in Japan resulted in the release of radioactive materials into the atmosphere, the nearby sea, and the surrounding land. Following the accident, several meteorological models were used to predict the transport of the radioactive materials to other continents such as North America and Europe. Also of high importance is the dispersion of radioactive materials locally and within Japan. Based on the International Atomic Energy Agency (IAEA) Convention on Early Notification of a nuclear accident, several radiological data sets were collected on the accident by the Japanese authorities. Among the radioactive materials monitored, are I-131more » and Cs-137 which form the major contributions to the contamination of drinking water. The radiation dose in the atmosphere was also measured. It is impractical to measure contamination and radiation dose in every place of interest. Therefore, modeling helps to predict contamination and radiation dose. Some modeling studies that have been reported in the literature include the simulation of transport and deposition of I-131 and Cs-137 from the accident, Cs-137 deposition and contamination of Japanese soils, and preliminary estimates of I-131 and Cs-137 discharged from the plant into the atmosphere. In this paper, we present statistical analytics of I-131 and Cs-137 with the goal of predicting gamma dose from the Fukushima Daiichi nuclear accident. The data sets used in our study were collected from the IAEA Fukushima Monitoring Database. As part of this study, we investigated several regression models to find the best algorithm for modeling the gamma dose. The modeling techniques used in our study include linear regression, principal component regression (PCR), partial least square (PLS) regression, and ridge regression. Our preliminary results on the first set of data showed that the linear regression model with one variable was the best with a root mean square error of 0.0133 μSv/h, compared to 0.0210 μSv/h for PCR, 0.231 μSv/h for ridge regression L-curve, 0.0856 μSv/h for PLS, and 0.0860 μSv/h for ridge regression cross validation. Complete results using the full datasets for these models will also be presented. (authors)« less

Support vector regression methodology for estimating global solar radiation in Algeria

NASA Astrophysics Data System (ADS)

Guermoui, Mawloud; Rabehi, Abdelaziz; Gairaa, Kacem; Benkaciali, Said

2018-01-01

Accurate estimation of Daily Global Solar Radiation (DGSR) has been a major goal for solar energy applications. In this paper we show the possibility of developing a simple model based on the Support Vector Regression (SVM-R), which could be used to estimate DGSR on the horizontal surface in Algeria based only on sunshine ratio as input. The SVM model has been developed and tested using a data set recorded over three years (2005-2007). The data was collected at the Applied Research Unit for Renewable Energies (URAER) in Ghardaïa city. The data collected between 2005-2006 are used to train the model while the 2007 data are used to test the performance of the selected model. The measured and the estimated values of DGSR were compared during the testing phase statistically using the Root Mean Square Error (RMSE), Relative Square Error (rRMSE), and correlation coefficient (r2), which amount to 1.59(MJ/m2), 8.46 and 97,4%, respectively. The obtained results show that the SVM-R is highly qualified for DGSR estimation using only sunshine ratio.

Trees grow on money: urban tree canopy cover and environmental justice.

PubMed

Schwarz, Kirsten; Fragkias, Michail; Boone, Christopher G; Zhou, Weiqi; McHale, Melissa; Grove, J Morgan; O'Neil-Dunne, Jarlath; McFadden, Joseph P; Buckley, Geoffrey L; Childers, Dan; Ogden, Laura; Pincetl, Stephanie; Pataki, Diane; Whitmer, Ali; Cadenasso, Mary L

2015-01-01

This study examines the distributional equity of urban tree canopy (UTC) cover for Baltimore, MD, Los Angeles, CA, New York, NY, Philadelphia, PA, Raleigh, NC, Sacramento, CA, and Washington, D.C. using high spatial resolution land cover data and census data. Data are analyzed at the Census Block Group levels using Spearman's correlation, ordinary least squares regression (OLS), and a spatial autoregressive model (SAR). Across all cities there is a strong positive correlation between UTC cover and median household income. Negative correlations between race and UTC cover exist in bivariate models for some cities, but they are generally not observed using multivariate regressions that include additional variables on income, education, and housing age. SAR models result in higher r-square values compared to the OLS models across all cities, suggesting that spatial autocorrelation is an important feature of our data. Similarities among cities can be found based on shared characteristics of climate, race/ethnicity, and size. Our findings suggest that a suite of variables, including income, contribute to the distribution of UTC cover. These findings can help target simultaneous strategies for UTC goals and environmental justice concerns.

Kernel analysis of partial least squares (PLS) regression models.

PubMed

Shinzawa, Hideyuki; Ritthiruangdej, Pitiporn; Ozaki, Yukihiro

2011-05-01

An analytical technique based on kernel matrix representation is demonstrated to provide further chemically meaningful insight into partial least squares (PLS) regression models. The kernel matrix condenses essential information about scores derived from PLS or principal component analysis (PCA). Thus, it becomes possible to establish the proper interpretation of the scores. A PLS model for the total nitrogen (TN) content in multiple Thai fish sauces is built with a set of near-infrared (NIR) transmittance spectra of the fish sauce samples. The kernel analysis of the scores effectively reveals that the variation of the spectral feature induced by the change in protein content is substantially associated with the total water content and the protein hydration. Kernel analysis is also carried out on a set of time-dependent infrared (IR) spectra representing transient evaporation of ethanol from a binary mixture solution of ethanol and oleic acid. A PLS model to predict the elapsed time is built with the IR spectra and the kernel matrix is derived from the scores. The detailed analysis of the kernel matrix provides penetrating insight into the interaction between the ethanol and the oleic acid.

Prediction of octanol-water partition coefficients of organic compounds by multiple linear regression, partial least squares, and artificial neural network.

PubMed

Golmohammadi, Hassan

2009-11-30

A quantitative structure-property relationship (QSPR) study was performed to develop models those relate the structure of 141 organic compounds to their octanol-water partition coefficients (log P(o/w)). A genetic algorithm was applied as a variable selection tool. Modeling of log P(o/w) of these compounds as a function of theoretically derived descriptors was established by multiple linear regression (MLR), partial least squares (PLS), and artificial neural network (ANN). The best selected descriptors that appear in the models are: atomic charge weighted partial positively charged surface area (PPSA-3), fractional atomic charge weighted partial positive surface area (FPSA-3), minimum atomic partial charge (Qmin), molecular volume (MV), total dipole moment of molecule (mu), maximum antibonding contribution of a molecule orbital in the molecule (MAC), and maximum free valency of a C atom in the molecule (MFV). The result obtained showed the ability of developed artificial neural network to prediction of partition coefficients of organic compounds. Also, the results revealed the superiority of ANN over the MLR and PLS models. Copyright 2009 Wiley Periodicals, Inc.

An Application to the Prediction of LOD Change Based on General Regression Neural Network

NASA Astrophysics Data System (ADS)

Zhang, X. H.; Wang, Q. J.; Zhu, J. J.; Zhang, H.

2011-07-01

Traditional prediction of the LOD (length of day) change was based on linear models, such as the least square model and the autoregressive technique, etc. Due to the complex non-linear features of the LOD variation, the performances of the linear model predictors are not fully satisfactory. This paper applies a non-linear neural network - general regression neural network (GRNN) model to forecast the LOD change, and the results are analyzed and compared with those obtained with the back propagation neural network and other models. The comparison shows that the performance of the GRNN model in the prediction of the LOD change is efficient and feasible.

An Alternative Method for Allocating Base Maintenance Supplies to Mission, Design, and Series Aircraft in the United States Air Force.

DTIC Science & Technology

1987-09-01

Edition,. Fail 1986. 33. Neter, John et al. Applied Linear Regression MoceL. Homewood IL: Richard D. Irwin, Incorporated, iJ83. 34. NovicK, David... Linear Regression Models (33) then, for each sample observation (X fh, the method of least squares considers the deviation of Yubms from its expected value...for finding good estimators of b - b5 * In -2raer to explain the procedure, the model Yubms = b0 + b!xfh will be discussed. According to Applied

A method for nonlinear exponential regression analysis

NASA Technical Reports Server (NTRS)

Junkin, B. G.

1971-01-01

A computer-oriented technique is presented for performing a nonlinear exponential regression analysis on decay-type experimental data. The technique involves the least squares procedure wherein the nonlinear problem is linearized by expansion in a Taylor series. A linear curve fitting procedure for determining the initial nominal estimates for the unknown exponential model parameters is included as an integral part of the technique. A correction matrix was derived and then applied to the nominal estimate to produce an improved set of model parameters. The solution cycle is repeated until some predetermined criterion is satisfied.

Least Square Fast Learning Network for modeling the combustion efficiency of a 300WM coal-fired boiler.

PubMed

Li, Guoqiang; Niu, Peifeng; Wang, Huaibao; Liu, Yongchao

2014-03-01

This paper presents a novel artificial neural network with a very fast learning speed, all of whose weights and biases are determined by the twice Least Square method, so it is called Least Square Fast Learning Network (LSFLN). In addition, there is another difference from conventional neural networks, which is that the output neurons of LSFLN not only receive the information from the hidden layer neurons, but also receive the external information itself directly from the input neurons. In order to test the validity of LSFLN, it is applied to 6 classical regression applications, and also employed to build the functional relation between the combustion efficiency and operating parameters of a 300WM coal-fired boiler. Experimental results show that, compared with other methods, LSFLN with very less hidden neurons could achieve much better regression precision and generalization ability at a much faster learning speed. Copyright © 2013 Elsevier Ltd. All rights reserved.

[Prediction of total nitrogen and alkali hydrolysable nitrogen content in loess using hyperspectral data based on correlation analysis and partial least squares regression].

PubMed

Liu, Xiu-ying; Wang, Li; Chang, Qing-rui; Wang, Xiao-xing; Shang, Yan

2015-07-01

Wuqi County of Shaanxi Province, where the vegetation recovering measures have been carried out for years, was taken as the study area. A total of 100 loess samples from 24 different profiles were collected. Total nitrogen (TN) and alkali hydrolysable nitrogen (AHN) contents of the soil samples were analyzed, and the soil samples were scanned in the visible/near-infrared (VNIR) region of 350-2500 nm in the laboratory. The calibration models were developed between TN and AHN contents and VNIR values based on correlation analysis (CA) and partial least squares regression (PLS). Independent samples validated the calibration models. The results indicated that the optimum model for predicting TN of loess was established by using first derivative of reflectance. The best model for predicting AHN of loess was established by using normal derivative spectra. The optimum TN model could effectively predict TN in loess from 0 to 40 cm, but the optimum AHN model could only roughly predict AHN at the same depth. This study provided a good method for rapidly predicting TN of loess where vegetation recovering measures have been adopted, but prediction of AHN needs to be further studied.

Time series modeling by a regression approach based on a latent process.

PubMed

Chamroukhi, Faicel; Samé, Allou; Govaert, Gérard; Aknin, Patrice

2009-01-01

Time series are used in many domains including finance, engineering, economics and bioinformatics generally to represent the change of a measurement over time. Modeling techniques may then be used to give a synthetic representation of such data. A new approach for time series modeling is proposed in this paper. It consists of a regression model incorporating a discrete hidden logistic process allowing for activating smoothly or abruptly different polynomial regression models. The model parameters are estimated by the maximum likelihood method performed by a dedicated Expectation Maximization (EM) algorithm. The M step of the EM algorithm uses a multi-class Iterative Reweighted Least-Squares (IRLS) algorithm to estimate the hidden process parameters. To evaluate the proposed approach, an experimental study on simulated data and real world data was performed using two alternative approaches: a heteroskedastic piecewise regression model using a global optimization algorithm based on dynamic programming, and a Hidden Markov Regression Model whose parameters are estimated by the Baum-Welch algorithm. Finally, in the context of the remote monitoring of components of the French railway infrastructure, and more particularly the switch mechanism, the proposed approach has been applied to modeling and classifying time series representing the condition measurements acquired during switch operations.

Application of least median of squared orthogonal distance (LMD) and LMD-based reweighted least squares (RLS) methods on the stock-recruitment relationship

NASA Astrophysics Data System (ADS)

Wang, Yan-Jun; Liu, Qun

1999-03-01

Analysis of stock-recruitment (SR) data is most often done by fitting various SR relationship curves to the data. Fish population dynamics data often have stochastic variations and measurement errors, which usually result in a biased regression analysis. This paper presents a robust regression method, least median of squared orthogonal distance (LMD), which is insensitive to abnormal values in the dependent and independent variables in a regression analysis. Outliers that have significantly different variance from the rest of the data can be identified in a residual analysis. Then, the least squares (LS) method is applied to the SR data with defined outliers being down weighted. The application of LMD and LMD-based Reweighted Least Squares (RLS) method to simulated and real fisheries SR data is explored.

A Weighted Least Squares Approach To Robustify Least Squares Estimates.

ERIC Educational Resources Information Center

Lin, Chowhong; Davenport, Ernest C., Jr.

This study developed a robust linear regression technique based on the idea of weighted least squares. In this technique, a subsample of the full data of interest is drawn, based on a measure of distance, and an initial set of regression coefficients is calculated. The rest of the data points are then taken into the subsample, one after another,…

Validation of Core Temperature Estimation Algorithm

DTIC Science & Technology

2016-01-29

plot of observed versus estimated core temperature with the line of identity (dashed) and the least squares regression line (solid) and line equation...estimated PSI with the line of identity (dashed) and the least squares regression line (solid) and line equation in the top left corner. (b) Bland...for comparison. The root mean squared error (RMSE) was also computed, as given by Equation 2.

Quantile regression applied to spectral distance decay

USGS Publications Warehouse

Rocchini, D.; Cade, B.S.

2008-01-01

Remotely sensed imagery has long been recognized as a powerful support for characterizing and estimating biodiversity. Spectral distance among sites has proven to be a powerful approach for detecting species composition variability. Regression analysis of species similarity versus spectral distance allows us to quantitatively estimate the amount of turnover in species composition with respect to spectral and ecological variability. In classical regression analysis, the residual sum of squares is minimized for the mean of the dependent variable distribution. However, many ecological data sets are characterized by a high number of zeroes that add noise to the regression model. Quantile regressions can be used to evaluate trend in the upper quantiles rather than a mean trend across the whole distribution of the dependent variable. In this letter, we used ordinary least squares (OLS) and quantile regressions to estimate the decay of species similarity versus spectral distance. The achieved decay rates were statistically nonzero (p < 0.01), considering both OLS and quantile regressions. Nonetheless, the OLS regression estimate of the mean decay rate was only half the decay rate indicated by the upper quantiles. Moreover, the intercept value, representing the similarity reached when the spectral distance approaches zero, was very low compared with the intercepts of the upper quantiles, which detected high species similarity when habitats are more similar. In this letter, we demonstrated the power of using quantile regressions applied to spectral distance decay to reveal species diversity patterns otherwise lost or underestimated by OLS regression. ?? 2008 IEEE.

Estimation of Flood-Frequency Discharges for Rural, Unregulated Streams in West Virginia

USGS Publications Warehouse

Wiley, Jeffrey B.; Atkins, John T.

2010-01-01

Flood-frequency discharges were determined for 290 streamgage stations having a minimum of 9 years of record in West Virginia and surrounding states through the 2006 or 2007 water year. No trend was determined in the annual peaks used to calculate the flood-frequency discharges. Multiple and simple least-squares regression equations for the 100-year (1-percent annual-occurrence probability) flood discharge with independent variables that describe the basin characteristics were developed for 290 streamgage stations in West Virginia and adjacent states. The regression residuals for the models were evaluated and used to define three regions of the State, designated as Eastern Panhandle, Central Mountains, and Western Plateaus. Exploratory data analysis procedures identified 44 streamgage stations that were excluded from the development of regression equations representative of rural, unregulated streams in West Virginia. Regional equations for the 1.1-, 1.5-, 2-, 5-, 10-, 25-, 50-, 100-, 200-, and 500-year flood discharges were determined by generalized least-squares regression using data from the remaining 246 streamgage stations. Drainage area was the only significant independent variable determined for all equations in all regions. Procedures developed to estimate flood-frequency discharges on ungaged streams were based on (1) regional equations and (2) drainage-area ratios between gaged and ungaged locations on the same stream. The procedures are applicable only to rural, unregulated streams within the boundaries of West Virginia that have drainage areas within the limits of the stations used to develop the regional equations (from 0.21 to 1,461 square miles in the Eastern Panhandle, from 0.10 to 1,619 square miles in the Central Mountains, and from 0.13 to 1,516 square miles in the Western Plateaus). The accuracy of the equations is quantified by measuring the average prediction error (from 21.7 to 56.3 percent) and equivalent years of record (from 2.0 to 70.9 years).

An Analysis of San Diego's Housing Market Using a Geographically Weighted Regression Approach

NASA Astrophysics Data System (ADS)

Grant, Christina P.

San Diego County real estate transaction data was evaluated with a set of linear models calibrated by ordinary least squares and geographically weighted regression (GWR). The goal of the analysis was to determine whether the spatial effects assumed to be in the data are best studied globally with no spatial terms, globally with a fixed effects submarket variable, or locally with GWR. 18,050 single-family residential sales which closed in the six months between April 2014 and September 2014 were used in the analysis. Diagnostic statistics including AICc, R2, Global Moran's I, and visual inspection of diagnostic plots and maps indicate superior model performance by GWR as compared to both global regressions.

Landslide susceptibility mapping using decision-tree based CHi-squared automatic interaction detection (CHAID) and Logistic regression (LR) integration

NASA Astrophysics Data System (ADS)

Althuwaynee, Omar F.; Pradhan, Biswajeet; Ahmad, Noordin

2014-06-01

This article uses methodology based on chi-squared automatic interaction detection (CHAID), as a multivariate method that has an automatic classification capacity to analyse large numbers of landslide conditioning factors. This new algorithm was developed to overcome the subjectivity of the manual categorization of scale data of landslide conditioning factors, and to predict rainfall-induced susceptibility map in Kuala Lumpur city and surrounding areas using geographic information system (GIS). The main objective of this article is to use CHi-squared automatic interaction detection (CHAID) method to perform the best classification fit for each conditioning factor, then, combining it with logistic regression (LR). LR model was used to find the corresponding coefficients of best fitting function that assess the optimal terminal nodes. A cluster pattern of landslide locations was extracted in previous study using nearest neighbor index (NNI), which were then used to identify the clustered landslide locations range. Clustered locations were used as model training data with 14 landslide conditioning factors such as; topographic derived parameters, lithology, NDVI, land use and land cover maps. Pearson chi-squared value was used to find the best classification fit between the dependent variable and conditioning factors. Finally the relationship between conditioning factors were assessed and the landslide susceptibility map (LSM) was produced. An area under the curve (AUC) was used to test the model reliability and prediction capability with the training and validation landslide locations respectively. This study proved the efficiency and reliability of decision tree (DT) model in landslide susceptibility mapping. Also it provided a valuable scientific basis for spatial decision making in planning and urban management studies.

[Application of near infrared spectroscopy combined with particle swarm optimization based least square support vactor machine to rapid quantitative analysis of Corni Fructus].

PubMed

Liu, Xue-song; Sun, Fen-fang; Jin, Ye; Wu, Yong-jiang; Gu, Zhi-xin; Zhu, Li; Yan, Dong-lan

2015-12-01

A novel method was developed for the rapid determination of multi-indicators in corni fructus by means of near infrared (NIR) spectroscopy. Particle swarm optimization (PSO) based least squares support vector machine was investigated to increase the levels of quality control. The calibration models of moisture, extractum, morroniside and loganin were established using the PSO-LS-SVM algorithm. The performance of PSO-LS-SVM models was compared with partial least squares regression (PLSR) and back propagation artificial neural network (BP-ANN). The calibration and validation results of PSO-LS-SVM were superior to both PLS and BP-ANN. For PSO-LS-SVM models, the correlation coefficients (r) of calibrations were all above 0.942. The optimal prediction results were also achieved by PSO-LS-SVM models with the RMSEP (root mean square error of prediction) and RSEP (relative standard errors of prediction) less than 1.176 and 15.5% respectively. The results suggest that PSO-LS-SVM algorithm has a good model performance and high prediction accuracy. NIR has a potential value for rapid determination of multi-indicators in Corni Fructus.

Comparing least-squares and quantile regression approaches to analyzing median hospital charges.

PubMed

Olsen, Cody S; Clark, Amy E; Thomas, Andrea M; Cook, Lawrence J

2012-07-01

Emergency department (ED) and hospital charges obtained from administrative data sets are useful descriptors of injury severity and the burden to EDs and the health care system. However, charges are typically positively skewed due to costly procedures, long hospital stays, and complicated or prolonged treatment for few patients. The median is not affected by extreme observations and is useful in describing and comparing distributions of hospital charges. A least-squares analysis employing a log transformation is one approach for estimating median hospital charges, corresponding confidence intervals (CIs), and differences between groups; however, this method requires certain distributional properties. An alternate method is quantile regression, which allows estimation and inference related to the median without making distributional assumptions. The objective was to compare the log-transformation least-squares method to the quantile regression approach for estimating median hospital charges, differences in median charges between groups, and associated CIs. The authors performed simulations using repeated sampling of observed statewide ED and hospital charges and charges randomly generated from a hypothetical lognormal distribution. The median and 95% CI and the multiplicative difference between the median charges of two groups were estimated using both least-squares and quantile regression methods. Performance of the two methods was evaluated. In contrast to least squares, quantile regression produced estimates that were unbiased and had smaller mean square errors in simulations of observed ED and hospital charges. Both methods performed well in simulations of hypothetical charges that met least-squares method assumptions. When the data did not follow the assumed distribution, least-squares estimates were often biased, and the associated CIs had lower than expected coverage as sample size increased. Quantile regression analyses of hospital charges provide unbiased estimates even when lognormal and equal variance assumptions are violated. These methods may be particularly useful in describing and analyzing hospital charges from administrative data sets. © 2012 by the Society for Academic Emergency Medicine.

Water Quality Variable Estimation using Partial Least Squares Regression and Multi-Scale Remote Sensing.

NASA Astrophysics Data System (ADS)

Peterson, K. T.; Wulamu, A.

2017-12-01

Water, essential to all living organisms, is one of the Earth's most precious resources. Remote sensing offers an ideal approach to monitor water quality over traditional in-situ techniques that are highly time and resource consuming. Utilizing a multi-scale approach, incorporating data from handheld spectroscopy, UAS based hyperspectal, and satellite multispectral images were collected in coordination with in-situ water quality samples for the two midwestern watersheds. The remote sensing data was modeled and correlated to the in-situ water quality variables including chlorophyll content (Chl), turbidity, and total dissolved solids (TDS) using Normalized Difference Spectral Indices (NDSI) and Partial Least Squares Regression (PLSR). The results of the study supported the original hypothesis that correlating water quality variables with remotely sensed data benefits greatly from the use of more complex modeling and regression techniques such as PLSR. The final results generated from the PLSR analysis resulted in much higher R2 values for all variables when compared to NDSI. The combination of NDSI and PLSR analysis also identified key wavelengths for identification that aligned with previous study's findings. This research displays the advantages and future for complex modeling and machine learning techniques to improve water quality variable estimation from spectral data.

Providing the Fire Risk Map in Forest Area Using a Geographically Weighted Regression Model with Gaussin Kernel and Modis Images, a Case Study: Golestan Province

NASA Astrophysics Data System (ADS)

Shah-Heydari pour, A.; Pahlavani, P.; Bigdeli, B.

2017-09-01

According to the industrialization of cities and the apparent increase in pollutants and greenhouse gases, the importance of forests as the natural lungs of the earth is felt more than ever to clean these pollutants. Annually, a large part of the forests is destroyed due to the lack of timely action during the fire. Knowledge about areas with a high-risk of fire and equipping these areas by constructing access routes and allocating the fire-fighting equipment can help to eliminate the destruction of the forest. In this research, the fire risk of region was forecasted and the risk map of that was provided using MODIS images by applying geographically weighted regression model with Gaussian kernel and ordinary least squares over the effective parameters in forest fire including distance from residential areas, distance from the river, distance from the road, height, slope, aspect, soil type, land use, average temperature, wind speed, and rainfall. After the evaluation, it was found that the geographically weighted regression model with Gaussian kernel forecasted 93.4% of the all fire points properly, however the ordinary least squares method could forecast properly only 66% of the fire points.

Development of regional skews for selected flood durations for the Central Valley Region, California, based on data through water year 2008

USGS Publications Warehouse

Lamontagne, Jonathan R.; Stedinger, Jery R.; Berenbrock, Charles; Veilleux, Andrea G.; Ferris, Justin C.; Knifong, Donna L.

2012-01-01

Flood-frequency information is important in the Central Valley region of California because of the high risk of catastrophic flooding. Most traditional flood-frequency studies focus on peak flows, but for the assessment of the adequacy of reservoirs, levees, other flood control structures, sustained flood flow (flood duration) frequency data are needed. This study focuses on rainfall or rain-on-snow floods, rather than the annual maximum, because rain events produce the largest floods in the region. A key to estimating flood-duration frequency is determining the regional skew for such data. Of the 50 sites used in this study to determine regional skew, 28 sites were considered to have little to no significant regulated flows, and for the 22 sites considered significantly regulated, unregulated daily flow data were synthesized by using reservoir storage changes and diversion records. The unregulated, annual maximum rainfall flood flows for selected durations (1-day, 3-day, 7-day, 15-day, and 30-day) for all 50 sites were furnished by the U.S. Army Corps of Engineers. Station skew was determined by using the expected moments algorithm program for fitting the Pearson Type 3 flood-frequency distribution to the logarithms of annual flood-duration data. Bayesian generalized least squares regression procedures used in earlier studies were modified to address problems caused by large cross correlations among concurrent rainfall floods in California and to address the extensive censoring of low outliers at some sites, by using the new expected moments algorithm for fitting the LP3 distribution to rainfall flood-duration data. To properly account for these problems and to develop suitable regional-skew regression models and regression diagnostics, a combination of ordinary least squares, weighted least squares, and Bayesian generalized least squares regressions were adopted. This new methodology determined that a nonlinear model relating regional skew to mean basin elevation was the best model for each flood duration. The regional-skew values ranged from -0.74 for a flood duration of 1-day and a mean basin elevation less than 2,500 feet to values near 0 for a flood duration of 7-days and a mean basin elevation greater than 4,500 feet. This relation between skew and elevation reflects the interaction of snow and rain, which increases with increased elevation. The regional skews are more accurate, and the mean squared errors are less than in the Interagency Advisory Committee on Water Data's National skew map of Bulletin 17B.

Least median of squares and iteratively re-weighted least squares as robust linear regression methods for fluorimetric determination of α-lipoic acid in capsules in ideal and non-ideal cases of linearity.

PubMed

Korany, Mohamed A; Gazy, Azza A; Khamis, Essam F; Ragab, Marwa A A; Kamal, Miranda F

2018-06-01

This study outlines two robust regression approaches, namely least median of squares (LMS) and iteratively re-weighted least squares (IRLS) to investigate their application in instrument analysis of nutraceuticals (that is, fluorescence quenching of merbromin reagent upon lipoic acid addition). These robust regression methods were used to calculate calibration data from the fluorescence quenching reaction (∆F and F-ratio) under ideal or non-ideal linearity conditions. For each condition, data were treated using three regression fittings: Ordinary Least Squares (OLS), LMS and IRLS. Assessment of linearity, limits of detection (LOD) and quantitation (LOQ), accuracy and precision were carefully studied for each condition. LMS and IRLS regression line fittings showed significant improvement in correlation coefficients and all regression parameters for both methods and both conditions. In the ideal linearity condition, the intercept and slope changed insignificantly, but a dramatic change was observed for the non-ideal condition and linearity intercept. Under both linearity conditions, LOD and LOQ values after the robust regression line fitting of data were lower than those obtained before data treatment. The results obtained after statistical treatment indicated that the linearity ranges for drug determination could be expanded to lower limits of quantitation by enhancing the regression equation parameters after data treatment. Analysis results for lipoic acid in capsules, using both fluorimetric methods, treated by parametric OLS and after treatment by robust LMS and IRLS were compared for both linearity conditions. Copyright © 2018 John Wiley & Sons, Ltd.

Estimating standard errors in feature network models.

PubMed

Frank, Laurence E; Heiser, Willem J

2007-05-01

Feature network models are graphical structures that represent proximity data in a discrete space while using the same formalism that is the basis of least squares methods employed in multidimensional scaling. Existing methods to derive a network model from empirical data only give the best-fitting network and yield no standard errors for the parameter estimates. The additivity properties of networks make it possible to consider the model as a univariate (multiple) linear regression problem with positivity restrictions on the parameters. In the present study, both theoretical and empirical standard errors are obtained for the constrained regression parameters of a network model with known features. The performance of both types of standard error is evaluated using Monte Carlo techniques.

Blood glucose prediction using neural network

NASA Astrophysics Data System (ADS)

Soh, Chit Siang; Zhang, Xiqin; Chen, Jianhong; Raveendran, P.; Soh, Phey Hong; Yeo, Joon Hock

2008-02-01

We used neural network for blood glucose level determination in this study. The data set used in this study was collected using a non-invasive blood glucose monitoring system with six laser diodes, each laser diode operating at distinct near infrared wavelength between 1500nm and 1800nm. The neural network is specifically used to determine blood glucose level of one individual who participated in an oral glucose tolerance test (OGTT) session. Partial least squares regression is also used for blood glucose level determination for the purpose of comparison with the neural network model. The neural network model performs better in the prediction of blood glucose level as compared with the partial least squares model.

Combined computational-experimental approach to predict blood-brain barrier (BBB) permeation based on "green" salting-out thin layer chromatography supported by simple molecular descriptors.

PubMed

Ciura, Krzesimir; Belka, Mariusz; Kawczak, Piotr; Bączek, Tomasz; Markuszewski, Michał J; Nowakowska, Joanna

2017-09-05

The objective of this paper is to build QSRR/QSAR model for predicting the blood-brain barrier (BBB) permeability. The obtained models are based on salting-out thin layer chromatography (SOTLC) constants and calculated molecular descriptors. Among chromatographic methods SOTLC was chosen, since the mobile phases are free of organic solvent. As consequences, there are less toxic, and have lower environmental impact compared to classical reserved phases liquid chromatography (RPLC). During the study three stationary phase silica gel, cellulose plates and neutral aluminum oxide were examined. The model set of solutes presents a wide range of log BB values, containing compounds which cross the BBB readily and molecules poorly distributed to the brain including drugs acting on the nervous system as well as peripheral acting drugs. Additionally, the comparison of three regression models: multiple linear regression (MLR), partial least-squares (PLS) and orthogonal partial least squares (OPLS) were performed. The designed QSRR/QSAR models could be useful to predict BBB of systematically synthesized newly compounds in the drug development pipeline and are attractive alternatives of time-consuming and demanding directed methods for log BB measurement. The study also shown that among several regression techniques, significant differences can be obtained in models performance, measured by R 2 and Q 2 , hence it is strongly suggested to evaluate all available options as MLR, PLS and OPLS. Copyright © 2017 Elsevier B.V. All rights reserved.

Bitterness intensity prediction of berberine hydrochloride using an electronic tongue and a GA-BP neural network.

PubMed

Liu, Ruixin; Zhang, Xiaodong; Zhang, Lu; Gao, Xiaojie; Li, Huiling; Shi, Junhan; Li, Xuelin

2014-06-01

The aim of this study was to predict the bitterness intensity of a drug using an electronic tongue (e-tongue). The model drug of berberine hydrochloride was used to establish a bitterness prediction model (BPM), based on the taste evaluation of bitterness intensity by a taste panel, the data provided by the e-tongue and a genetic algorithm-back-propagation neural network (GA-BP) modeling method. The modeling characteristics of the GA-BP were compared with those of multiple linear regression, partial least square regression and BP methods. The determination coefficient of the BPM was 0.99965±0.00004, the root mean square error of cross-validation was 0.1398±0.0488 and the correlation coefficient of the cross-validation between the true and predicted values was 0.9959±0.0027. The model is superior to the other three models based on these indicators. In conclusion, the model established in this study has a high fitting degree and may be used for the bitterness prediction modeling of berberine hydrochloride of different concentrations. The model also provides a reference for the generation of BPMs of other drugs. Additionally, the algorithm of the study is able to conduct a rapid and accurate quantitative analysis of the data provided by the e-tongue.

Bitterness intensity prediction of berberine hydrochloride using an electronic tongue and a GA-BP neural network

PubMed Central

LIU, RUIXIN; ZHANG, XIAODONG; ZHANG, LU; GAO, XIAOJIE; LI, HUILING; SHI, JUNHAN; LI, XUELIN

2014-01-01

The aim of this study was to predict the bitterness intensity of a drug using an electronic tongue (e-tongue). The model drug of berberine hydrochloride was used to establish a bitterness prediction model (BPM), based on the taste evaluation of bitterness intensity by a taste panel, the data provided by the e-tongue and a genetic algorithm-back-propagation neural network (GA-BP) modeling method. The modeling characteristics of the GA-BP were compared with those of multiple linear regression, partial least square regression and BP methods. The determination coefficient of the BPM was 0.99965±0.00004, the root mean square error of cross-validation was 0.1398±0.0488 and the correlation coefficient of the cross-validation between the true and predicted values was 0.9959±0.0027. The model is superior to the other three models based on these indicators. In conclusion, the model established in this study has a high fitting degree and may be used for the bitterness prediction modeling of berberine hydrochloride of different concentrations. The model also provides a reference for the generation of BPMs of other drugs. Additionally, the algorithm of the study is able to conduct a rapid and accurate quantitative analysis of the data provided by the e-tongue. PMID:24926369

Correlation of sensory bitterness in dairy protein hydrolysates: Comparison of prediction models built using sensory, chromatographic and electronic tongue data.

PubMed

Newman, J; Egan, T; Harbourne, N; O'Riordan, D; Jacquier, J C; O'Sullivan, M

2014-08-01

Sensory evaluation can be problematic for ingredients with a bitter taste during research and development phase of new food products. In this study, 19 dairy protein hydrolysates (DPH) were analysed by an electronic tongue and their physicochemical characteristics, the data obtained from these methods were correlated with their bitterness intensity as scored by a trained sensory panel and each model was also assessed by its predictive capabilities. The physiochemical characteristics of the DPHs investigated were degree of hydrolysis (DH%), and data relating to peptide size and relative hydrophobicity from size exclusion chromatography (SEC) and reverse phase (RP) HPLC. Partial least square regression (PLS) was used to construct the prediction models. All PLS regressions had good correlations (0.78 to 0.93) with the strongest being the combination of data obtained from SEC and RP HPLC. However, the PLS with the strongest predictive power was based on the e-tongue which had the PLS regression with the lowest root mean predicted residual error sum of squares (PRESS) in the study. The results show that the PLS models constructed with the e-tongue and the combination of SEC and RP-HPLC has potential to be used for prediction of bitterness and thus reducing the reliance on sensory analysis in DPHs for future food research. Copyright © 2014 Elsevier B.V. All rights reserved.

SPSS macros to compare any two fitted values from a regression model.

PubMed

Weaver, Bruce; Dubois, Sacha

2012-12-01

In regression models with first-order terms only, the coefficient for a given variable is typically interpreted as the change in the fitted value of Y for a one-unit increase in that variable, with all other variables held constant. Therefore, each regression coefficient represents the difference between two fitted values of Y. But the coefficients represent only a fraction of the possible fitted value comparisons that might be of interest to researchers. For many fitted value comparisons that are not captured by any of the regression coefficients, common statistical software packages do not provide the standard errors needed to compute confidence intervals or carry out statistical tests-particularly in more complex models that include interactions, polynomial terms, or regression splines. We describe two SPSS macros that implement a matrix algebra method for comparing any two fitted values from a regression model. The !OLScomp and !MLEcomp macros are for use with models fitted via ordinary least squares and maximum likelihood estimation, respectively. The output from the macros includes the standard error of the difference between the two fitted values, a 95% confidence interval for the difference, and a corresponding statistical test with its p-value.

Application of General Regression Neural Network to the Prediction of LOD Change

NASA Astrophysics Data System (ADS)

Zhang, Xiao-Hong; Wang, Qi-Jie; Zhu, Jian-Jun; Zhang, Hao

2012-01-01

Traditional methods for predicting the change in length of day (LOD change) are mainly based on some linear models, such as the least square model and autoregression model, etc. However, the LOD change comprises complicated non-linear factors and the prediction effect of the linear models is always not so ideal. Thus, a kind of non-linear neural network — general regression neural network (GRNN) model is tried to make the prediction of the LOD change and the result is compared with the predicted results obtained by taking advantage of the BP (back propagation) neural network model and other models. The comparison result shows that the application of the GRNN to the prediction of the LOD change is highly effective and feasible.

An open-access CMIP5 pattern library for temperature and precipitation: Description and methodology

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

Lynch, Cary D.; Hartin, Corinne A.; Bond-Lamberty, Benjamin

Pattern scaling is used to efficiently emulate general circulation models and explore uncertainty in climate projections under multiple forcing scenarios. Pattern scaling methods assume that local climate changes scale with a global mean temperature increase, allowing for spatial patterns to be generated for multiple models for any future emission scenario. For uncertainty quantification and probabilistic statistical analysis, a library of patterns with descriptive statistics for each file would be beneficial, but such a library does not presently exist. Of the possible techniques used to generate patterns, the two most prominent are the delta and least squared regression methods. We exploremore » the differences and statistical significance between patterns generated by each method and assess performance of the generated patterns across methods and scenarios. Differences in patterns across seasons between methods and epochs were largest in high latitudes (60-90°N/S). Bias and mean errors between modeled and pattern predicted output from the linear regression method were smaller than patterns generated by the delta method. Across scenarios, differences in the linear regression method patterns were more statistically significant, especially at high latitudes. We found that pattern generation methodologies were able to approximate the forced signal of change to within ≤ 0.5°C, but choice of pattern generation methodology for pattern scaling purposes should be informed by user goals and criteria. As a result, this paper describes our library of least squared regression patterns from all CMIP5 models for temperature and precipitation on an annual and sub-annual basis, along with the code used to generate these patterns.« less

An open-access CMIP5 pattern library for temperature and precipitation: Description and methodology

DOE PAGES

Lynch, Cary D.; Hartin, Corinne A.; Bond-Lamberty, Benjamin; ...

2017-05-15

Pattern scaling is used to efficiently emulate general circulation models and explore uncertainty in climate projections under multiple forcing scenarios. Pattern scaling methods assume that local climate changes scale with a global mean temperature increase, allowing for spatial patterns to be generated for multiple models for any future emission scenario. For uncertainty quantification and probabilistic statistical analysis, a library of patterns with descriptive statistics for each file would be beneficial, but such a library does not presently exist. Of the possible techniques used to generate patterns, the two most prominent are the delta and least squared regression methods. We exploremore » the differences and statistical significance between patterns generated by each method and assess performance of the generated patterns across methods and scenarios. Differences in patterns across seasons between methods and epochs were largest in high latitudes (60-90°N/S). Bias and mean errors between modeled and pattern predicted output from the linear regression method were smaller than patterns generated by the delta method. Across scenarios, differences in the linear regression method patterns were more statistically significant, especially at high latitudes. We found that pattern generation methodologies were able to approximate the forced signal of change to within ≤ 0.5°C, but choice of pattern generation methodology for pattern scaling purposes should be informed by user goals and criteria. As a result, this paper describes our library of least squared regression patterns from all CMIP5 models for temperature and precipitation on an annual and sub-annual basis, along with the code used to generate these patterns.« less

Partial Least Squares Regression Can Aid in Detecting Differential Abundance of Multiple Features in Sets of Metagenomic Samples

PubMed Central

Libiger, Ondrej; Schork, Nicholas J.

2015-01-01

It is now feasible to examine the composition and diversity of microbial communities (i.e., “microbiomes”) that populate different human organs and orifices using DNA sequencing and related technologies. To explore the potential links between changes in microbial communities and various diseases in the human body, it is essential to test associations involving different species within and across microbiomes, environmental settings and disease states. Although a number of statistical techniques exist for carrying out relevant analyses, it is unclear which of these techniques exhibit the greatest statistical power to detect associations given the complexity of most microbiome datasets. We compared the statistical power of principal component regression, partial least squares regression, regularized regression, distance-based regression, Hill's diversity measures, and a modified test implemented in the popular and widely used microbiome analysis methodology “Metastats” across a wide range of simulated scenarios involving changes in feature abundance between two sets of metagenomic samples. For this purpose, simulation studies were used to change the abundance of microbial species in a real dataset from a published study examining human hands. Each technique was applied to the same data, and its ability to detect the simulated change in abundance was assessed. We hypothesized that a small subset of methods would outperform the rest in terms of the statistical power. Indeed, we found that the Metastats technique modified to accommodate multivariate analysis and partial least squares regression yielded high power under the models and data sets we studied. The statistical power of diversity measure-based tests, distance-based regression and regularized regression was significantly lower. Our results provide insight into powerful analysis strategies that utilize information on species counts from large microbiome data sets exhibiting skewed frequency distributions obtained on a small to moderate number of samples. PMID:26734061

The influencing factors of CO2 emission intensity of Chinese agriculture from 1997 to 2014.

PubMed

Long, Xingle; Luo, Yusen; Wu, Chao; Zhang, Jijian

2018-05-01

In China, agriculture produces the greatest chemical oxygen demand (COD) emissions in wastewater and the most methane (CH 4 ) emissions. It is imperative that agricultural pollution in China be reduced. This study investigated the influencing factors of the CO 2 emission intensity of Chinese agriculture from 1997 to 2014. We analyzed the influencing factors of the CO 2 emission intensity through the first-stage least-square regression. We also analyzed determinants of innovation through the second-stage least-square regression. We found that innovation negatively affected the CO 2 emission intensity in the model of the nation. FDI positively affected innovation in China. It is important to enhance indigenous innovation for green agriculture through labor training and collaboration between agriculture and academia.

The Detection and Quantification of Adulteration in Ground Roasted Asian Palm Civet Coffee Using Near-Infrared Spectroscopy in Tandem with Chemometrics

NASA Astrophysics Data System (ADS)

Suhandy, D.; Yulia, M.; Ogawa, Y.; Kondo, N.

2018-05-01

In the present research, an evaluation of using near infrared (NIR) spectroscopy in tandem with full spectrum partial least squares (FS-PLS) regression for quantification of degree of adulteration in civet coffee was conducted. A number of 126 ground roasted coffee samples with degree of adulteration 0-51% were prepared. Spectral data were acquired using a NIR spectrometer equipped with an integrating sphere for diffuse reflectance measurement in the range of 1300-2500 nm. The samples were divided into two groups calibration sample set (84 samples) and prediction sample set (42 samples). The calibration model was developed on original spectra using FS-PLS regression with full-cross validation method. The calibration model exhibited the determination coefficient R2=0.96 for calibration and R2=0.92 for validation. The prediction resulted in low root mean square error of prediction (RMSEP) (4.67%) and high ratio prediction to deviation (RPD) (3.75). In conclusion, the degree of adulteration in civet coffee have been quantified successfully by using NIR spectroscopy and FS-PLS regression in a non-destructive, economical, precise, and highly sensitive method, which uses very simple sample preparation.

Development of Super-Ensemble techniques for ocean analyses: the Mediterranean Sea case

NASA Astrophysics Data System (ADS)

Pistoia, Jenny; Pinardi, Nadia; Oddo, Paolo; Collins, Matthew; Korres, Gerasimos; Drillet, Yann

2017-04-01

Short-term ocean analyses for Sea Surface Temperature SST in the Mediterranean Sea can be improved by a statistical post-processing technique, called super-ensemble. This technique consists in a multi-linear regression algorithm applied to a Multi-Physics Multi-Model Super-Ensemble (MMSE) dataset, a collection of different operational forecasting analyses together with ad-hoc simulations produced by modifying selected numerical model parameterizations. A new linear regression algorithm based on Empirical Orthogonal Function filtering techniques is capable to prevent overfitting problems, even if best performances are achieved when we add correlation to the super-ensemble structure using a simple spatial filter applied after the linear regression. Our outcomes show that super-ensemble performances depend on the selection of an unbiased operator and the length of the learning period, but the quality of the generating MMSE dataset has the largest impact on the MMSE analysis Root Mean Square Error (RMSE) evaluated with respect to observed satellite SST. Lower RMSE analysis estimates result from the following choices: 15 days training period, an overconfident MMSE dataset (a subset with the higher quality ensemble members), and the least square algorithm being filtered a posteriori.

Predicting recreational water quality advisories: A comparison of statistical methods

USGS Publications Warehouse

Brooks, Wesley R.; Corsi, Steven R.; Fienen, Michael N.; Carvin, Rebecca B.

2016-01-01

Epidemiological studies indicate that fecal indicator bacteria (FIB) in beach water are associated with illnesses among people having contact with the water. In order to mitigate public health impacts, many beaches are posted with an advisory when the concentration of FIB exceeds a beach action value. The most commonly used method of measuring FIB concentration takes 18–24 h before returning a result. In order to avoid the 24 h lag, it has become common to ”nowcast” the FIB concentration using statistical regressions on environmental surrogate variables. Most commonly, nowcast models are estimated using ordinary least squares regression, but other regression methods from the statistical and machine learning literature are sometimes used. This study compares 14 regression methods across 7 Wisconsin beaches to identify which consistently produces the most accurate predictions. A random forest model is identified as the most accurate, followed by multiple regression fit using the adaptive LASSO.

A Bayesian model averaging method for the derivation of reservoir operating rules

NASA Astrophysics Data System (ADS)

Zhang, Jingwen; Liu, Pan; Wang, Hao; Lei, Xiaohui; Zhou, Yanlai

2015-09-01

Because the intrinsic dynamics among optimal decision making, inflow processes and reservoir characteristics are complex, functional forms of reservoir operating rules are always determined subjectively. As a result, the uncertainty of selecting form and/or model involved in reservoir operating rules must be analyzed and evaluated. In this study, we analyze the uncertainty of reservoir operating rules using the Bayesian model averaging (BMA) model. Three popular operating rules, namely piecewise linear regression, surface fitting and a least-squares support vector machine, are established based on the optimal deterministic reservoir operation. These individual models provide three-member decisions for the BMA combination, enabling the 90% release interval to be estimated by the Markov Chain Monte Carlo simulation. A case study of China's the Baise reservoir shows that: (1) the optimal deterministic reservoir operation, superior to any reservoir operating rules, is used as the samples to derive the rules; (2) the least-squares support vector machine model is more effective than both piecewise linear regression and surface fitting; (3) BMA outperforms any individual model of operating rules based on the optimal trajectories. It is revealed that the proposed model can reduce the uncertainty of operating rules, which is of great potential benefit in evaluating the confidence interval of decisions.

An operational GLS model for hydrologic regression

USGS Publications Warehouse

Tasker, Gary D.; Stedinger, J.R.

1989-01-01

Recent Monte Carlo studies have documented the value of generalized least squares (GLS) procedures to estimate empirical relationships between streamflow statistics and physiographic basin characteristics. This paper presents a number of extensions of the GLS method that deal with realities and complexities of regional hydrologic data sets that were not addressed in the simulation studies. These extensions include: (1) a more realistic model of the underlying model errors; (2) smoothed estimates of cross correlation of flows; (3) procedures for including historical flow data; (4) diagnostic statistics describing leverage and influence for GLS regression; and (5) the formulation of a mathematical program for evaluating future gaging activities. ?? 1989.

Obscure phenomena in statistical analysis of quantitative structure-activity relationships. Part 1: Multicollinearity of physicochemical descriptors.

PubMed

Mager, P P; Rothe, H

1990-10-01

Multicollinearity of physicochemical descriptors leads to serious consequences in quantitative structure-activity relationship (QSAR) analysis, such as incorrect estimators and test statistics of regression coefficients of the ordinary least-squares (OLS) model applied usually to QSARs. Beside the diagnosis of the known simple collinearity, principal component regression analysis (PCRA) also allows the diagnosis of various types of multicollinearity. Only if the absolute values of PCRA estimators are order statistics that decrease monotonically, the effects of multicollinearity can be circumvented. Otherwise, obscure phenomena may be observed, such as good data recognition but low predictive model power of a QSAR model.

Estimating magnitude and frequency of peak discharges for rural, unregulated, streams in West Virginia

USGS Publications Warehouse

Wiley, J.B.; Atkins, John T.; Tasker, Gary D.

2000-01-01

Multiple and simple least-squares regression models for the log10-transformed 100-year discharge with independent variables describing the basin characteristics (log10-transformed and untransformed) for 267 streamflow-gaging stations were evaluated, and the regression residuals were plotted as areal distributions that defined three regions of the State, designated East, North, and South. Exploratory data analysis procedures identified 31 gaging stations at which discharges are different than would be expected for West Virginia. Regional equations for the 2-, 5-, 10-, 25-, 50-, 100-, 200-, and 500-year peak discharges were determined by generalized least-squares regression using data from 236 gaging stations. Log10-transformed drainage area was the most significant independent variable for all regions.Equations developed in this study are applicable only to rural, unregulated, streams within the boundaries of West Virginia. The accuracy of estimating equations is quantified by measuring the average prediction error (from 27.7 to 44.7 percent) and equivalent years of record (from 1.6 to 20.0 years).

Regression Models and Fuzzy Logic Prediction of TBM Penetration Rate

NASA Astrophysics Data System (ADS)

Minh, Vu Trieu; Katushin, Dmitri; Antonov, Maksim; Veinthal, Renno

2017-03-01

This paper presents statistical analyses of rock engineering properties and the measured penetration rate of tunnel boring machine (TBM) based on the data of an actual project. The aim of this study is to analyze the influence of rock engineering properties including uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), rock brittleness index (BI), the distance between planes of weakness (DPW), and the alpha angle (Alpha) between the tunnel axis and the planes of weakness on the TBM rate of penetration (ROP). Four (4) statistical regression models (two linear and two nonlinear) are built to predict the ROP of TBM. Finally a fuzzy logic model is developed as an alternative method and compared to the four statistical regression models. Results show that the fuzzy logic model provides better estimations and can be applied to predict the TBM performance. The R-squared value (R2) of the fuzzy logic model scores the highest value of 0.714 over the second runner-up of 0.667 from the multiple variables nonlinear regression model.

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

Experimental evaluation of four infiltration models for calcareous soil irrigated with treated untreated grey water and fresh water

NASA Astrophysics Data System (ADS)

Gharaibeh, M. A.; Eltaif, N. I.; Alrababah, M. A.; Alhamad, M. N.

2009-04-01

Infiltration is vital for both irrigated and rainfed agriculture. The knowledge of infiltration characteristics of a soil is the basic information required for designing an efficient irrigation system. The objective of the present study was to model soil infiltration using four models: Green and Ampt, Horton, Kostaikov and modified Kostiakov. Infiltration tests were conducted on field plot irrigated with treated, untreated greywater and fresh water. The field water infiltration data used in these models were based on double ring infiltrometer tests conducted for 4 h. The algebraic parameters of the infiltration models and nonlinear least squares regression were fitted using measured infiltration time [I (t)] data. Among process-based infiltration models, the Horton model performed best and matched the measured I (t) data with lower sum of squares (SS).

A robust nonparametric framework for reconstruction of stochastic differential equation models

NASA Astrophysics Data System (ADS)

Rajabzadeh, Yalda; Rezaie, Amir Hossein; Amindavar, Hamidreza

2016-05-01

In this paper, we employ a nonparametric framework to robustly estimate the functional forms of drift and diffusion terms from discrete stationary time series. The proposed method significantly improves the accuracy of the parameter estimation. In this framework, drift and diffusion coefficients are modeled through orthogonal Legendre polynomials. We employ the least squares regression approach along with the Euler-Maruyama approximation method to learn coefficients of stochastic model. Next, a numerical discrete construction of mean squared prediction error (MSPE) is established to calculate the order of Legendre polynomials in drift and diffusion terms. We show numerically that the new method is robust against the variation in sample size and sampling rate. The performance of our method in comparison with the kernel-based regression (KBR) method is demonstrated through simulation and real data. In case of real dataset, we test our method for discriminating healthy electroencephalogram (EEG) signals from epilepsy ones. We also demonstrate the efficiency of the method through prediction in the financial data. In both simulation and real data, our algorithm outperforms the KBR method.

Prediction of olive oil sensory descriptors using instrumental data fusion and partial least squares (PLS) regression.

PubMed

Borràs, Eva; Ferré, Joan; Boqué, Ricard; Mestres, Montserrat; Aceña, Laura; Calvo, Angels; Busto, Olga

2016-08-01

Headspace-Mass Spectrometry (HS-MS), Fourier Transform Mid-Infrared spectroscopy (FT-MIR) and UV-Visible spectrophotometry (UV-vis) instrumental responses have been combined to predict virgin olive oil sensory descriptors. 343 olive oil samples analyzed during four consecutive harvests (2010-2014) were used to build multivariate calibration models using partial least squares (PLS) regression. The reference values of the sensory attributes were provided by expert assessors from an official taste panel. The instrumental data were modeled individually and also using data fusion approaches. The use of fused data with both low- and mid-level of abstraction improved PLS predictions for all the olive oil descriptors. The best PLS models were obtained for two positive attributes (fruity and bitter) and two defective descriptors (fusty and musty), all of them using data fusion of MS and MIR spectral fingerprints. Although good predictions were not obtained for some sensory descriptors, the results are encouraging, specially considering that the legal categorization of virgin olive oils only requires the determination of fruity and defective descriptors. Copyright © 2016 Elsevier B.V. All rights reserved.

Use of Vis/NIRS for the determination of sugar content of cola soft drinks based on chemometric methods

NASA Astrophysics Data System (ADS)

Liu, Fei; He, Yong

2008-03-01

Three different chemometric methods were performed for the determination of sugar content of cola soft drinks using visible and near infrared spectroscopy (Vis/NIRS). Four varieties of colas were prepared and 180 samples (45 samples for each variety) were selected for the calibration set, while 60 samples (15 samples for each variety) for the validation set. The smoothing way of Savitzky-Golay, standard normal variate (SNV) and Savitzky-Golay first derivative transformation were applied for the pre-processing of spectral data. The first eleven principal components (PCs) extracted by partial least squares (PLS) analysis were employed as the inputs of BP neural network (BPNN) and least squares-support vector machine (LS-SVM) model. Then the BPNN model with the optimal structural parameters and LS-SVM model with radial basis function (RBF) kernel were applied to build the regression model with a comparison of PLS regression. The correlation coefficient (r), root mean square error of prediction (RMSEP) and bias for prediction were 0.971, 1.259 and -0.335 for PLS, 0.986, 0.763, and -0.042 for BPNN, while 0.978, 0.995 and -0.227 for LS-SVM, respectively. All the three methods supplied a high and satisfying precision. The results indicated that Vis/NIR spectroscopy combined with chemometric methods could be utilized as a high precision way for the determination of sugar content of cola soft drinks.

Quantitative analysis of bayberry juice acidity based on visible and near-infrared spectroscopy

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

Shao Yongni; He Yong; Mao Jingyuan

Visible and near-infrared (Vis/NIR) reflectance spectroscopy has been investigated for its ability to nondestructively detect acidity in bayberry juice. What we believe to be a new, better mathematic model is put forward, which we have named principal component analysis-stepwise regression analysis-backpropagation neural network (PCA-SRA-BPNN), to build a correlation between the spectral reflectivity data and the acidity of bayberry juice. In this model, the optimum network parameters,such as the number of input nodes, hidden nodes, learning rate, and momentum, are chosen by the value of root-mean-square (rms) error. The results show that its prediction statistical parameters are correlation coefficient (r) ofmore » 0.9451 and root-mean-square error of prediction(RMSEP) of 0.1168. Partial least-squares (PLS) regression is also established to compare with this model. Before doing this, the influences of various spectral pretreatments (standard normal variate, multiplicative scatter correction, S. Golay first derivative, and wavelet package transform) are compared. The PLS approach with wavelet package transform preprocessing spectra is found to provide the best results, and its prediction statistical parameters are correlation coefficient (r) of 0.9061 and RMSEP of 0.1564. Hence, these two models are both desirable to analyze the data from Vis/NIR spectroscopy and to solve the problem of the acidity prediction of bayberry juice. This supplies basal research to ultimately realize the online measurements of the juice's internal quality through this Vis/NIR spectroscopy technique.« less

Functional Data Analysis Applied to Modeling of Severe Acute Mucositis and Dysphagia Resulting From Head and Neck Radiation Therapy.

PubMed

Dean, Jamie A; Wong, Kee H; Gay, Hiram; Welsh, Liam C; Jones, Ann-Britt; Schick, Ulrike; Oh, Jung Hun; Apte, Aditya; Newbold, Kate L; Bhide, Shreerang A; Harrington, Kevin J; Deasy, Joseph O; Nutting, Christopher M; Gulliford, Sarah L

2016-11-15

Current normal tissue complication probability modeling using logistic regression suffers from bias and high uncertainty in the presence of highly correlated radiation therapy (RT) dose data. This hinders robust estimates of dose-response associations and, hence, optimal normal tissue-sparing strategies from being elucidated. Using functional data analysis (FDA) to reduce the dimensionality of the dose data could overcome this limitation. FDA was applied to modeling of severe acute mucositis and dysphagia resulting from head and neck RT. Functional partial least squares regression (FPLS) and functional principal component analysis were used for dimensionality reduction of the dose-volume histogram data. The reduced dose data were input into functional logistic regression models (functional partial least squares-logistic regression [FPLS-LR] and functional principal component-logistic regression [FPC-LR]) along with clinical data. This approach was compared with penalized logistic regression (PLR) in terms of predictive performance and the significance of treatment covariate-response associations, assessed using bootstrapping. The area under the receiver operating characteristic curve for the PLR, FPC-LR, and FPLS-LR models was 0.65, 0.69, and 0.67, respectively, for mucositis (internal validation) and 0.81, 0.83, and 0.83, respectively, for dysphagia (external validation). The calibration slopes/intercepts for the PLR, FPC-LR, and FPLS-LR models were 1.6/-0.67, 0.45/0.47, and 0.40/0.49, respectively, for mucositis (internal validation) and 2.5/-0.96, 0.79/-0.04, and 0.79/0.00, respectively, for dysphagia (external validation). The bootstrapped odds ratios indicated significant associations between RT dose and severe toxicity in the mucositis and dysphagia FDA models. Cisplatin was significantly associated with severe dysphagia in the FDA models. None of the covariates was significantly associated with severe toxicity in the PLR models. Dose levels greater than approximately 1.0 Gy/fraction were most strongly associated with severe acute mucositis and dysphagia in the FDA models. FPLS and functional principal component analysis marginally improved predictive performance compared with PLR and provided robust dose-response associations. FDA is recommended for use in normal tissue complication probability modeling. Copyright © 2016 The Author(s). Published by Elsevier Inc. All rights reserved.

Institutional Focus and Non-Resident Student Enrollment

ERIC Educational Resources Information Center

Baryla, Edward A., Jr.; Dotterweich, Douglas

2006-01-01

Purpose: This paper uses institutional characteristics and regional economic data to determine if institutional mission may help drive non-resident undergraduate enrollment. Design/methodology/approach: A two-stage least squares regression models is employed on 180 Doctoral, 333 Comprehensive, and 501 Baccalaureate higher education institutions to…

Random forest models to predict aqueous solubility.

PubMed

Palmer, David S; O'Boyle, Noel M; Glen, Robert C; Mitchell, John B O

2007-01-01

Random Forest regression (RF), Partial-Least-Squares (PLS) regression, Support Vector Machines (SVM), and Artificial Neural Networks (ANN) were used to develop QSPR models for the prediction of aqueous solubility, based on experimental data for 988 organic molecules. The Random Forest regression model predicted aqueous solubility more accurately than those created by PLS, SVM, and ANN and offered methods for automatic descriptor selection, an assessment of descriptor importance, and an in-parallel measure of predictive ability, all of which serve to recommend its use. The prediction of log molar solubility for an external test set of 330 molecules that are solid at 25 degrees C gave an r2 = 0.89 and RMSE = 0.69 log S units. For a standard data set selected from the literature, the model performed well with respect to other documented methods. Finally, the diversity of the training and test sets are compared to the chemical space occupied by molecules in the MDL drug data report, on the basis of molecular descriptors selected by the regression analysis.

External characteristic determination of eggs and cracked eggs identification using spectral signature

PubMed Central

Xie, Chuanqi; He, Yong

2016-01-01

This study was carried out to use hyperspectral imaging technique for determining color (L*, a* and b*) and eggshell strength and identifying cracked chicken eggs. Partial least squares (PLS) models based on full and selected wavelengths suggested by regression coefficient (RC) method were established to predict the four parameters, respectively. Partial least squares-discriminant analysis (PLS-DA) and RC-partial least squares-discriminant analysis (RC-PLS-DA) models were applied to identify cracked eggs. PLS models performed well with the correlation coefficient (rp) of 0.788 for L*, 0.810 for a*, 0.766 for b* and 0.835 for eggshell strength. RC-PLS models also obtained the rp of 0.771 for L*, 0.806 for a*, 0.767 for b* and 0.841 for eggshell strength. The classification results were 97.06% in PLS-DA model and 88.24% in RC-PLS-DA model. It demonstrated that hyperspectral imaging technique has the potential to be used to detect color and eggshell strength values and identify cracked chicken eggs. PMID:26882990

Modeling groundwater nitrate concentrations in private wells in Iowa

USGS Publications Warehouse

Wheeler, David C.; Nolan, Bernard T.; Flory, Abigail R.; DellaValle, Curt T.; Ward, Mary H.

2015-01-01

Contamination of drinking water by nitrate is a growing problem in many agricultural areas of the country. Ingested nitrate can lead to the endogenous formation of N-nitroso compounds, potent carcinogens. We developed a predictive model for nitrate concentrations in private wells in Iowa. Using 34,084 measurements of nitrate in private wells, we trained and tested random forest models to predict log nitrate levels by systematically assessing the predictive performance of 179 variables in 36 thematic groups (well depth, distance to sinkholes, location, land use, soil characteristics, nitrogen inputs, meteorology, and other factors). The final model contained 66 variables in 17 groups. Some of the most important variables were well depth, slope length within 1 km of the well, year of sample, and distance to nearest animal feeding operation. The correlation between observed and estimated nitrate concentrations was excellent in the training set (r-square = 0.77) and was acceptable in the testing set (r-square = 0.38). The random forest model had substantially better predictive performance than a traditional linear regression model or a regression tree. Our model will be used to investigate the association between nitrate levels in drinking water and cancer risk in the Iowa participants of the Agricultural Health Study cohort.

Modeling groundwater nitrate concentrations in private wells in Iowa.

PubMed

Wheeler, David C; Nolan, Bernard T; Flory, Abigail R; DellaValle, Curt T; Ward, Mary H

2015-12-01

Contamination of drinking water by nitrate is a growing problem in many agricultural areas of the country. Ingested nitrate can lead to the endogenous formation of N-nitroso compounds, potent carcinogens. We developed a predictive model for nitrate concentrations in private wells in Iowa. Using 34,084 measurements of nitrate in private wells, we trained and tested random forest models to predict log nitrate levels by systematically assessing the predictive performance of 179 variables in 36 thematic groups (well depth, distance to sinkholes, location, land use, soil characteristics, nitrogen inputs, meteorology, and other factors). The final model contained 66 variables in 17 groups. Some of the most important variables were well depth, slope length within 1 km of the well, year of sample, and distance to nearest animal feeding operation. The correlation between observed and estimated nitrate concentrations was excellent in the training set (r-square=0.77) and was acceptable in the testing set (r-square=0.38). The random forest model had substantially better predictive performance than a traditional linear regression model or a regression tree. Our model will be used to investigate the association between nitrate levels in drinking water and cancer risk in the Iowa participants of the Agricultural Health Study cohort. Copyright © 2015 Elsevier B.V. All rights reserved.

Quantitative prediction of ionization effect on human skin permeability.

PubMed

Baba, Hiromi; Ueno, Yusuke; Hashida, Mitsuru; Yamashita, Fumiyoshi

2017-04-30

Although skin permeability of an active ingredient can be severely affected by its ionization in a dose solution, most of the existing prediction models cannot predict such impacts. To provide reliable predictors, we curated a novel large dataset of in vitro human skin permeability coefficients for 322 entries comprising chemically diverse permeants whose ionization fractions can be calculated. Subsequently, we generated thousands of computational descriptors, including LogD (octanol-water distribution coefficient at a specific pH), and analyzed the dataset using nonlinear support vector regression (SVR) and Gaussian process regression (GPR) combined with greedy descriptor selection. The SVR model was slightly superior to the GPR model, with externally validated squared correlation coefficient, root mean square error, and mean absolute error values of 0.94, 0.29, and 0.21, respectively. These models indicate that Log D is effective for a comprehensive prediction of ionization effects on skin permeability. In addition, the proposed models satisfied the statistical criteria endorsed in recent model validation studies. These models can evaluate virtually generated compounds at any pH; therefore, they can be used for high-throughput evaluations of numerous active ingredients and optimization of their skin permeability with respect to permeant ionization. Copyright © 2017 Elsevier B.V. All rights reserved.

Simultaneous determination of estrogens (ethinylestradiol and norgestimate) concentrations in human and bovine serum albumin by use of fluorescence spectroscopy and multivariate regression analysis.

PubMed

Hordge, LaQuana N; McDaniel, Kiara L; Jones, Derick D; Fakayode, Sayo O

2016-05-15

The endocrine disruption property of estrogens necessitates the immediate need for effective monitoring and development of analytical protocols for their analyses in biological and human specimens. This study explores the first combined utility of a steady-state fluorescence spectroscopy and multivariate partial-least-square (PLS) regression analysis for the simultaneous determination of two estrogens (17α-ethinylestradiol (EE) and norgestimate (NOR)) concentrations in bovine serum albumin (BSA) and human serum albumin (HSA) samples. The influence of EE and NOR concentrations and temperature on the emission spectra of EE-HSA EE-BSA, NOR-HSA, and NOR-BSA complexes was also investigated. The binding of EE with HSA and BSA resulted in increase in emission characteristics of HSA and BSA and a significant blue spectra shift. In contrast, the interaction of NOR with HSA and BSA quenched the emission characteristics of HSA and BSA. The observed emission spectral shifts preclude the effective use of traditional univariate regression analysis of fluorescent data for the determination of EE and NOR concentrations in HSA and BSA samples. Multivariate partial-least-squares (PLS) regression analysis was utilized to correlate the changes in emission spectra with EE and NOR concentrations in HSA and BSA samples. The figures-of-merit of the developed PLS regression models were excellent, with limits of detection as low as 1.6×10(-8) M for EE and 2.4×10(-7) M for NOR and good linearity (R(2)>0.994985). The PLS models correctly predicted EE and NOR concentrations in independent validation HSA and BSA samples with a root-mean-square-percent-relative-error (RMS%RE) of less than 6.0% at physiological condition. On the contrary, the use of univariate regression resulted in poor predictions of EE and NOR in HSA and BSA samples, with RMS%RE larger than 40% at physiological conditions. High accuracy, low sensitivity, simplicity, low-cost with no prior analyte extraction or separation required makes this method promising, compelling, and attractive alternative for the rapid determination of estrogen concentrations in biomedical and biological specimens, pharmaceuticals, or environmental samples. Published by Elsevier B.V.

Developing and testing a global-scale regression model to quantify mean annual streamflow

NASA Astrophysics Data System (ADS)

Barbarossa, Valerio; Huijbregts, Mark A. J.; Hendriks, A. Jan; Beusen, Arthur H. W.; Clavreul, Julie; King, Henry; Schipper, Aafke M.

2017-01-01

Quantifying mean annual flow of rivers (MAF) at ungauged sites is essential for assessments of global water supply, ecosystem integrity and water footprints. MAF can be quantified with spatially explicit process-based models, which might be overly time-consuming and data-intensive for this purpose, or with empirical regression models that predict MAF based on climate and catchment characteristics. Yet, regression models have mostly been developed at a regional scale and the extent to which they can be extrapolated to other regions is not known. In this study, we developed a global-scale regression model for MAF based on a dataset unprecedented in size, using observations of discharge and catchment characteristics from 1885 catchments worldwide, measuring between 2 and 106 km2. In addition, we compared the performance of the regression model with the predictive ability of the spatially explicit global hydrological model PCR-GLOBWB by comparing results from both models to independent measurements. We obtained a regression model explaining 89% of the variance in MAF based on catchment area and catchment averaged mean annual precipitation and air temperature, slope and elevation. The regression model performed better than PCR-GLOBWB for the prediction of MAF, as root-mean-square error (RMSE) values were lower (0.29-0.38 compared to 0.49-0.57) and the modified index of agreement (d) was higher (0.80-0.83 compared to 0.72-0.75). Our regression model can be applied globally to estimate MAF at any point of the river network, thus providing a feasible alternative to spatially explicit process-based global hydrological models.

Regression Models of Quarterly Overhead Costs for Six Government Aerospace Contractors.

DTIC Science & Technology

1986-03-01

34 Testing ,, for Serial Correlation After Least Squares %Regression, Econometrica, Vol. 36, No. 1, pp. 133-150, January 1968. Intrili8ator M.D., Econometric ...to be superior. These two estimators are both two-stage estimators that are calculated utilizing Wallis’s test statistic for fourth-order...utilizing Wallis’s test statistic for fourth-order autocorrelation. NTIS C F’,& D tI1C T - .1 I -. . . ..- rJ ,. *p J • - DA 3

Regression-assisted deconvolution.

PubMed

McIntyre, Julie; Stefanski, Leonard A

2011-06-30

We present a semi-parametric deconvolution estimator for the density function of a random variable biX that is measured with error, a common challenge in many epidemiological studies. Traditional deconvolution estimators rely only on assumptions about the distribution of X and the error in its measurement, and ignore information available in auxiliary variables. Our method assumes the availability of a covariate vector statistically related to X by a mean-variance function regression model, where regression errors are normally distributed and independent of the measurement errors. Simulations suggest that the estimator achieves a much lower integrated squared error than the observed-data kernel density estimator when models are correctly specified and the assumption of normal regression errors is met. We illustrate the method using anthropometric measurements of newborns to estimate the density function of newborn length. Copyright © 2011 John Wiley & Sons, Ltd.

Can We Use Regression Modeling to Quantify Mean Annual Streamflow at a Global-Scale?

NASA Astrophysics Data System (ADS)

Barbarossa, V.; Huijbregts, M. A. J.; Hendriks, J. A.; Beusen, A.; Clavreul, J.; King, H.; Schipper, A.

2016-12-01

Quantifying mean annual flow of rivers (MAF) at ungauged sites is essential for a number of applications, including assessments of global water supply, ecosystem integrity and water footprints. MAF can be quantified with spatially explicit process-based models, which might be overly time-consuming and data-intensive for this purpose, or with empirical regression models that predict MAF based on climate and catchment characteristics. Yet, regression models have mostly been developed at a regional scale and the extent to which they can be extrapolated to other regions is not known. In this study, we developed a global-scale regression model for MAF using observations of discharge and catchment characteristics from 1,885 catchments worldwide, ranging from 2 to 106 km2 in size. In addition, we compared the performance of the regression model with the predictive ability of the spatially explicit global hydrological model PCR-GLOBWB [van Beek et al., 2011] by comparing results from both models to independent measurements. We obtained a regression model explaining 89% of the variance in MAF based on catchment area, mean annual precipitation and air temperature, average slope and elevation. The regression model performed better than PCR-GLOBWB for the prediction of MAF, as root-mean-square error values were lower (0.29 - 0.38 compared to 0.49 - 0.57) and the modified index of agreement was higher (0.80 - 0.83 compared to 0.72 - 0.75). Our regression model can be applied globally at any point of the river network, provided that the input parameters are within the range of values employed in the calibration of the model. The performance is reduced for water scarce regions and further research should focus on improving such an aspect for regression-based global hydrological models.

Detection and quantification of adulteration in sandalwood oil through near infrared spectroscopy.

PubMed

Kuriakose, Saji; Thankappan, Xavier; Joe, Hubert; Venkataraman, Venkateswaran

2010-10-01

The confirmation of authenticity of essential oils and the detection of adulteration are problems of increasing importance in the perfumes, pharmaceutical, flavor and fragrance industries. This is especially true for 'value added' products like sandalwood oil. A methodical study is conducted here to demonstrate the potential use of Near Infrared (NIR) spectroscopy along with multivariate calibration models like principal component regression (PCR) and partial least square regression (PLSR) as rapid analytical techniques for the qualitative and quantitative determination of adulterants in sandalwood oil. After suitable pre-processing of the NIR raw spectral data, the models are built-up by cross-validation. The lowest Root Mean Square Error of Cross-Validation and Calibration (RMSECV and RMSEC % v/v) are used as a decision supporting system to fix the optimal number of factors. The coefficient of determination (R(2)) and the Root Mean Square Error of Prediction (RMSEP % v/v) in the prediction sets are used as the evaluation parameters (R(2) = 0.9999 and RMSEP = 0.01355). The overall result leads to the conclusion that NIR spectroscopy with chemometric techniques could be successfully used as a rapid, simple, instant and non-destructive method for the detection of adulterants, even 1% of the low-grade oils, in the high quality form of sandalwood oil.

Synthesis of linear regression coefficients by recovering the within-study covariance matrix from summary statistics.

PubMed

Yoneoka, Daisuke; Henmi, Masayuki

2017-06-01

Recently, the number of regression models has dramatically increased in several academic fields. However, within the context of meta-analysis, synthesis methods for such models have not been developed in a commensurate trend. One of the difficulties hindering the development is the disparity in sets of covariates among literature models. If the sets of covariates differ across models, interpretation of coefficients will differ, thereby making it difficult to synthesize them. Moreover, previous synthesis methods for regression models, such as multivariate meta-analysis, often have problems because covariance matrix of coefficients (i.e. within-study correlations) or individual patient data are not necessarily available. This study, therefore, proposes a brief explanation regarding a method to synthesize linear regression models under different covariate sets by using a generalized least squares method involving bias correction terms. Especially, we also propose an approach to recover (at most) threecorrelations of covariates, which is required for the calculation of the bias term without individual patient data. Copyright © 2016 John Wiley & Sons, Ltd. Copyright © 2016 John Wiley & Sons, Ltd.

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.

A non-linear data mining parameter selection algorithm for continuous variables

PubMed Central

Razavi, Marianne; Brady, Sean

2017-01-01

In this article, we propose a new data mining algorithm, by which one can both capture the non-linearity in data and also find the best subset model. To produce an enhanced subset of the original variables, a preferred selection method should have the potential of adding a supplementary level of regression analysis that would capture complex relationships in the data via mathematical transformation of the predictors and exploration of synergistic effects of combined variables. The method that we present here has the potential to produce an optimal subset of variables, rendering the overall process of model selection more efficient. This algorithm introduces interpretable parameters by transforming the original inputs and also a faithful fit to the data. The core objective of this paper is to introduce a new estimation technique for the classical least square regression framework. This new automatic variable transformation and model selection method could offer an optimal and stable model that minimizes the mean square error and variability, while combining all possible subset selection methodology with the inclusion variable transformations and interactions. Moreover, this method controls multicollinearity, leading to an optimal set of explanatory variables. PMID:29131829

Modeling of feed-forward control using the partial least squares regression method in the tablet compression process.

PubMed

Hattori, Yusuke; Otsuka, Makoto

2017-05-30

In the pharmaceutical industry, the implementation of continuous manufacturing has been widely promoted in lieu of the traditional batch manufacturing approach. More specially, in recent years, the innovative concept of feed-forward control has been introduced in relation to process analytical technology. In the present study, we successfully developed a feed-forward control model for the tablet compression process by integrating data obtained from near-infrared (NIR) spectra and the physical properties of granules. In the pharmaceutical industry, batch manufacturing routinely allows for the preparation of granules with the desired properties through the manual control of process parameters. On the other hand, continuous manufacturing demands the automatic determination of these process parameters. Here, we proposed the development of a control model using the partial least squares regression (PLSR) method. The most significant feature of this method is the use of dataset integrating both the NIR spectra and the physical properties of the granules. Using our model, we determined that the properties of products, such as tablet weight and thickness, need to be included as independent variables in the PLSR analysis in order to predict unknown process parameters. Copyright © 2017 Elsevier B.V. All rights reserved.

Trees Grow on Money: Urban Tree Canopy Cover and Environmental Justice

PubMed Central

Schwarz, Kirsten; Fragkias, Michail; Boone, Christopher G.; Zhou, Weiqi; McHale, Melissa; Grove, J. Morgan; O’Neil-Dunne, Jarlath; McFadden, Joseph P.; Buckley, Geoffrey L.; Childers, Dan; Ogden, Laura; Pincetl, Stephanie; Pataki, Diane; Whitmer, Ali; Cadenasso, Mary L.

2015-01-01

This study examines the distributional equity of urban tree canopy (UTC) cover for Baltimore, MD, Los Angeles, CA, New York, NY, Philadelphia, PA, Raleigh, NC, Sacramento, CA, and Washington, D.C. using high spatial resolution land cover data and census data. Data are analyzed at the Census Block Group levels using Spearman’s correlation, ordinary least squares regression (OLS), and a spatial autoregressive model (SAR). Across all cities there is a strong positive correlation between UTC cover and median household income. Negative correlations between race and UTC cover exist in bivariate models for some cities, but they are generally not observed using multivariate regressions that include additional variables on income, education, and housing age. SAR models result in higher r-square values compared to the OLS models across all cities, suggesting that spatial autocorrelation is an important feature of our data. Similarities among cities can be found based on shared characteristics of climate, race/ethnicity, and size. Our findings suggest that a suite of variables, including income, contribute to the distribution of UTC cover. These findings can help target simultaneous strategies for UTC goals and environmental justice concerns. PMID:25830303

MERGANSER- Predicting Mercury Levels in Fish and Loons in New England Lakes

EPA Science Inventory

MERGANSER (MERcury Geo-spatial AssesmentS for the New England Region) is an empirical least squares multiple regression model using atmospheric deposition of mercury (Hg) and readily obtainable lake and watershed features to predict fish and common loon Hg (as methyl mercury) in ...

Partial Least Square Analyses of Landscape and Surface Water Biota Associations in the Savannah River Basin

EPA Science Inventory

Ecologists are often faced with problem of small sample size, correlated and large number of predictors, and high noise-to-signal relationships. This necessitates excluding important variables from the model when applying standard multiple or multivariate regression analyses. In ...

Bayesian Nonparametric Prediction and Statistical Inference

DTIC Science & Technology

1989-09-07

Kadane, J. (1980), "Bayesian decision theory and the sim- plification of models," in Evaluation of Econometric Models, J. Kmenta and J. Ramsey , eds...the random model and weighted least squares regression," in Evaluation of Econometric Models, ed. by J. Kmenta and J. Ramsey , Academic Press, 197-217...likelihood function. On the other hand, H. Jeffreys’s theory of hypothesis testing covers the most important situations in which the prior is not diffuse. See

Cost Estimation of Naval Ship Acquisition.

DTIC Science & Technology

1983-12-01

one a 9-sub- system model , the other a single total cost model . The models were developed using the linear least squares regression tech- nique with...to Linear Statistical Models , McGraw-Hill, 1961. 11. Helmer, F. T., Bibliography on Pricing Methodology and Cost Estimating, Dept. of Economics and...SUPPI.EMSaTARY NOTES IS. KWRo" (Cowaft. en tever aide of ..aesep M’ Idab~t 6 Week ONNa.) Cost estimation; Acquisition; Parametric cost estimate; linear

Poisson regression models outperform the geometrical model in estimating the peak-to-trough ratio of seasonal variation: a simulation study.

PubMed

Christensen, A L; Lundbye-Christensen, S; Dethlefsen, C

2011-12-01

Several statistical methods of assessing seasonal variation are available. Brookhart and Rothman [3] proposed a second-order moment-based estimator based on the geometrical model derived by Edwards [1], and reported that this estimator is superior in estimating the peak-to-trough ratio of seasonal variation compared with Edwards' estimator with respect to bias and mean squared error. Alternatively, seasonal variation may be modelled using a Poisson regression model, which provides flexibility in modelling the pattern of seasonal variation and adjustments for covariates. Based on a Monte Carlo simulation study three estimators, one based on the geometrical model, and two based on log-linear Poisson regression models, were evaluated in regards to bias and standard deviation (SD). We evaluated the estimators on data simulated according to schemes varying in seasonal variation and presence of a secular trend. All methods and analyses in this paper are available in the R package Peak2Trough[13]. Applying a Poisson regression model resulted in lower absolute bias and SD for data simulated according to the corresponding model assumptions. Poisson regression models had lower bias and SD for data simulated to deviate from the corresponding model assumptions than the geometrical model. This simulation study encourages the use of Poisson regression models in estimating the peak-to-trough ratio of seasonal variation as opposed to the geometrical model. Copyright © 2011 Elsevier Ireland Ltd. All rights reserved.

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.

Measuring the contribution of water and green space amenities to housing values: an application and comparison of spatially weighted hedonic models

Treesearch

Seong-Hoon Cho; J. Michael Bowker; William M. Park

2006-01-01

This study estimates the influence of proximity to water bodies and park amenities on residential housing values in Knox County, Tennessee, using the hedonic price approach. Values for proximity to water bodies and parks are first estimated globally with a standard ordinary least squares (OLS) model. A locally weighted regression model is then employed to investigate...

[Establishment of the Mathematical Model for PMI Estimation Using FTIR Spectroscopy and Data Mining Method].

PubMed

Wang, L; Qin, X C; Lin, H C; Deng, K F; Luo, Y W; Sun, Q R; Du, Q X; Wang, Z Y; Tuo, Y; Sun, J H

2018-02-01

To analyse the relationship between Fourier transform infrared （FTIR） spectrum of rat's spleen tissue and postmortem interval （PMI） for PMI estimation using FTIR spectroscopy combined with data mining method. Rats were sacrificed by cervical dislocation, and the cadavers were placed at 20 ℃. The FTIR spectrum data of rats' spleen tissues were taken and measured at different time points. After pretreatment, the data was analysed by data mining method. The absorption peak intensity of rat's spleen tissue spectrum changed with the PMI, while the absorption peak position was unchanged. The results of principal component analysis （PCA） showed that the cumulative contribution rate of the first three principal components was 96%. There was an obvious clustering tendency for the spectrum sample at each time point. The methods of partial least squares discriminant analysis （PLS-DA） and support vector machine classification （SVMC） effectively divided the spectrum samples with different PMI into four categories （0-24 h, 48-72 h, 96-120 h and 144-168 h）. The determination coefficient （ R ²） of the PMI estimation model established by PLS regression analysis was 0.96, and the root mean square error of calibration （RMSEC） and root mean square error of cross validation （RMSECV） were 9.90 h and 11.39 h respectively. In prediction set, the R ² was 0.97, and the root mean square error of prediction （RMSEP） was 10.49 h. The FTIR spectrum of the rat's spleen tissue can be effectively analyzed qualitatively and quantitatively by the combination of FTIR spectroscopy and data mining method, and the classification and PLS regression models can be established for PMI estimation. Copyright© by the Editorial Department of Journal of Forensic Medicine.

Four Major South Korea's Rivers Using Deep Learning Models.

PubMed

Lee, Sangmok; Lee, Donghyun

2018-06-24

Harmful algal blooms are an annual phenomenon that cause environmental damage, economic losses, and disease outbreaks. A fundamental solution to this problem is still lacking, thus, the best option for counteracting the effects of algal blooms is to improve advance warnings (predictions). However, existing physical prediction models have difficulties setting a clear coefficient indicating the relationship between each factor when predicting algal blooms, and many variable data sources are required for the analysis. These limitations are accompanied by high time and economic costs. Meanwhile, artificial intelligence and deep learning methods have become increasingly common in scientific research; attempts to apply the long short-term memory (LSTM) model to environmental research problems are increasing because the LSTM model exhibits good performance for time-series data prediction. However, few studies have applied deep learning models or LSTM to algal bloom prediction, especially in South Korea, where algal blooms occur annually. Therefore, we employed the LSTM model for algal bloom prediction in four major rivers of South Korea. We conducted short-term (one week) predictions by employing regression analysis and deep learning techniques on a newly constructed water quality and quantity dataset drawn from 16 dammed pools on the rivers. Three deep learning models (multilayer perceptron, MLP; recurrent neural network, RNN; and long short-term memory, LSTM) were used to predict chlorophyll-a, a recognized proxy for algal activity. The results were compared to those from OLS (ordinary least square) regression analysis and actual data based on the root mean square error (RSME). The LSTM model showed the highest prediction rate for harmful algal blooms and all deep learning models out-performed the OLS regression analysis. Our results reveal the potential for predicting algal blooms using LSTM and deep learning.

Rapid and simultaneous analysis of five alkaloids in four parts of Coptidis Rhizoma by near-infrared spectroscopy

NASA Astrophysics Data System (ADS)

Jintao, Xue; Yufei, Liu; Liming, Ye; Chunyan, Li; Quanwei, Yang; Weiying, Wang; Yun, Jing; Minxiang, Zhang; Peng, Li

2018-01-01

Near-Infrared Spectroscopy (NIRS) was first used to develop a method for rapid and simultaneous determination of 5 active alkaloids (berberine, coptisine, palmatine, epiberberine and jatrorrhizine) in 4 parts (rhizome, fibrous root, stem and leaf) of Coptidis Rhizoma. A total of 100 samples from 4 main places of origin were collected and studied. With HPLC analysis values as calibration reference, the quantitative analysis of 5 marker components was performed by two different modeling methods, partial least-squares (PLS) regression as linear regression and artificial neural networks (ANN) as non-linear regression. The results indicated that the 2 types of models established were robust, accurate and repeatable for five active alkaloids, and the ANN models was more suitable for the determination of berberine, coptisine and palmatine while the PLS model was more suitable for the analysis of epiberberine and jatrorrhizine. The performance of the optimal models was achieved as follows: the correlation coefficient (R) for berberine, coptisine, palmatine, epiberberine and jatrorrhizine was 0.9958, 0.9956, 0.9959, 0.9963 and 0.9923, respectively; the root mean square error of validation (RMSEP) was 0.5093, 0.0578, 0.0443, 0.0563 and 0.0090, respectively. Furthermore, for the comprehensive exploitation and utilization of plant resource of Coptidis Rhizoma, the established NIR models were used to analysis the content of 5 active alkaloids in 4 parts of Coptidis Rhizoma and 4 main origin of places. This work demonstrated that NIRS may be a promising method as routine screening for off-line fast analysis or on-line quality assessment of traditional Chinese medicine (TCM).

Application of Partial Least Square (PLS) Regression to Determine Landscape-Scale Aquatic Resources Vulnerability in the Ozark Mountains

EPA Science Inventory

Partial least squares (PLS) analysis offers a number of advantages over the more traditionally used regression analyses applied in landscape ecology, particularly for determining the associations among multiple constituents of surface water and landscape configuration. Common dat...

Orthogonalizing EM: A design-based least squares algorithm.

PubMed

Xiong, Shifeng; Dai, Bin; Huling, Jared; Qian, Peter Z G

We introduce an efficient iterative algorithm, intended for various least squares problems, based on a design of experiments perspective. The algorithm, called orthogonalizing EM (OEM), works for ordinary least squares and can be easily extended to penalized least squares. The main idea of the procedure is to orthogonalize a design matrix by adding new rows and then solve the original problem by embedding the augmented design in a missing data framework. We establish several attractive theoretical properties concerning OEM. For the ordinary least squares with a singular regression matrix, an OEM sequence converges to the Moore-Penrose generalized inverse-based least squares estimator. For ordinary and penalized least squares with various penalties, it converges to a point having grouping coherence for fully aliased regression matrices. Convergence and the convergence rate of the algorithm are examined. Finally, we demonstrate that OEM is highly efficient for large-scale least squares and penalized least squares problems, and is considerably faster than competing methods when n is much larger than p . Supplementary materials for this article are available online.

An open-access CMIP5 pattern library for temperature and precipitation: description and methodology

NASA Astrophysics Data System (ADS)

Lynch, Cary; Hartin, Corinne; Bond-Lamberty, Ben; Kravitz, Ben

2017-05-01

Pattern scaling is used to efficiently emulate general circulation models and explore uncertainty in climate projections under multiple forcing scenarios. Pattern scaling methods assume that local climate changes scale with a global mean temperature increase, allowing for spatial patterns to be generated for multiple models for any future emission scenario. For uncertainty quantification and probabilistic statistical analysis, a library of patterns with descriptive statistics for each file would be beneficial, but such a library does not presently exist. Of the possible techniques used to generate patterns, the two most prominent are the delta and least squares regression methods. We explore the differences and statistical significance between patterns generated by each method and assess performance of the generated patterns across methods and scenarios. Differences in patterns across seasons between methods and epochs were largest in high latitudes (60-90° N/S). Bias and mean errors between modeled and pattern-predicted output from the linear regression method were smaller than patterns generated by the delta method. Across scenarios, differences in the linear regression method patterns were more statistically significant, especially at high latitudes. We found that pattern generation methodologies were able to approximate the forced signal of change to within ≤ 0.5 °C, but the choice of pattern generation methodology for pattern scaling purposes should be informed by user goals and criteria. This paper describes our library of least squares regression patterns from all CMIP5 models for temperature and precipitation on an annual and sub-annual basis, along with the code used to generate these patterns. The dataset and netCDF data generation code are available at doi:10.5281/zenodo.495632.

Support vector machine regression (SVR/LS-SVM)--an alternative to neural networks (ANN) for analytical chemistry? Comparison of nonlinear methods on near infrared (NIR) spectroscopy data.

PubMed

Balabin, Roman M; Lomakina, Ekaterina I

2011-04-21

In this study, we make a general comparison of the accuracy and robustness of five multivariate calibration models: partial least squares (PLS) regression or projection to latent structures, polynomial partial least squares (Poly-PLS) regression, artificial neural networks (ANNs), and two novel techniques based on support vector machines (SVMs) for multivariate data analysis: support vector regression (SVR) and least-squares support vector machines (LS-SVMs). The comparison is based on fourteen (14) different datasets: seven sets of gasoline data (density, benzene content, and fractional composition/boiling points), two sets of ethanol gasoline fuel data (density and ethanol content), one set of diesel fuel data (total sulfur content), three sets of petroleum (crude oil) macromolecules data (weight percentages of asphaltenes, resins, and paraffins), and one set of petroleum resins data (resins content). Vibrational (near-infrared, NIR) spectroscopic data are used to predict the properties and quality coefficients of gasoline, biofuel/biodiesel, diesel fuel, and other samples of interest. The four systems presented here range greatly in composition, properties, strength of intermolecular interactions (e.g., van der Waals forces, H-bonds), colloid structure, and phase behavior. Due to the high diversity of chemical systems studied, general conclusions about SVM regression methods can be made. We try to answer the following question: to what extent can SVM-based techniques replace ANN-based approaches in real-world (industrial/scientific) applications? The results show that both SVR and LS-SVM methods are comparable to ANNs in accuracy. Due to the much higher robustness of the former, the SVM-based approaches are recommended for practical (industrial) application. This has been shown to be especially true for complicated, highly nonlinear objects.

Relative Motion of the WDS 05110+3203 STF 648 System, With a Protocol for Calculating Relative Motion

NASA Astrophysics Data System (ADS)

Wiley, E. O.

2010-07-01

Relative motion studies of visual double stars can be investigated using least squares regression techniques and readily accessible programs such as Microsoft Excel and a calculator. Optical pairs differ from physical pairs under most geometries in both their simple scatter plots and their regression models. A step-by-step protocol for estimating the rectilinear elements of an optical pair is presented. The characteristics of physical pairs using these techniques are discussed.

Computerized pigment design based on property hypersurfaces

NASA Astrophysics Data System (ADS)

Halova, Jaroslava; Sulcova, Petra; Kupka, Karel

2007-05-01

Competition is tough in the pigment market. Rational pigment design has therefore a competitive advantage, saving time and money. The aim of this work is to provide methods that can assist in designing pigments with defined properties. These methods include partial least squares regression (PLSR), neural network (NN) and generalized regression ANOVA model. Authors show how PLS bi-plot can be used to identify market gaps poorly covered by pigment manufacturers, thus giving an opportunity to develop pigments with potentially profitable properties.

Development of a Multiple Linear Regression Model to Forecast Facility Electrical Consumption at an Air Force Base.

DTIC Science & Technology

1981-09-01

corresponds to the same square footage that consumed the electrical energy. 3. The basic assumptions of multiple linear regres- sion, as enumerated in...7. Data related to the sample of bases is assumed to be representative of bases in the population. Limitations Basic limitations on this research were... Ratemaking --Overview. Rand Report R-5894, Santa Monica CA, May 1977. Chatterjee, Samprit, and Bertram Price. Regression Analysis by Example. New York: John

Near infrared spectrometric technique for testing fruit quality: optimisation of regression models using genetic algorithms

NASA Astrophysics Data System (ADS)

Isingizwe Nturambirwe, J. Frédéric; Perold, Willem J.; Opara, Umezuruike L.

2016-02-01

Near infrared (NIR) spectroscopy has gained extensive use in quality evaluation. It is arguably one of the most advanced spectroscopic tools in non-destructive quality testing of food stuff, from measurement to data analysis and interpretation. NIR spectral data are interpreted through means often involving multivariate statistical analysis, sometimes associated with optimisation techniques for model improvement. The objective of this research was to explore the extent to which genetic algorithms (GA) can be used to enhance model development, for predicting fruit quality. Apple fruits were used, and NIR spectra in the range from 12000 to 4000 cm-1 were acquired on both bruised and healthy tissues, with different degrees of mechanical damage. GAs were used in combination with partial least squares regression methods to develop bruise severity prediction models, and compared to PLS models developed using the full NIR spectrum. A classification model was developed, which clearly separated bruised from unbruised apple tissue. GAs helped improve prediction models by over 10%, in comparison with full spectrum-based models, as evaluated in terms of error of prediction (Root Mean Square Error of Cross-validation). PLS models to predict internal quality, such as sugar content and acidity were developed and compared to the versions optimized by genetic algorithm. Overall, the results highlighted the potential use of GA method to improve speed and accuracy of fruit quality prediction.

Early Maternal Employment and Children's School Readiness in Contemporary Families

ERIC Educational Resources Information Center

Lombardi, Caitlin McPherran; Coley, Rebekah Levine

2014-01-01

This study assessed whether previous findings linking early maternal employment to lower cognitive and behavioral skills among children generalized to modern families. Using a representative sample of children born in the United States in 2001 (N = 10,100), ordinary least squares regression models weighted with propensity scores assessed links…

Stressed and Losing Sleep: Sleep Duration and Perceived Stress among Affluent Adolescent Females

ERIC Educational Resources Information Center

DeSilva Mousseau, Angela M.; Lund, Terese J.; Liang, Belle; Spencer, Renée; Walsh, Jill

2016-01-01

This study examined the relationship between stress and sleep duration for adolescent females from affluent backgrounds. Participants were 218 students attending two independent single-sex secondary schools. Ordinary Least Squares (OLS) regression models (cross-sectional and longitudinal) were run to examine the association between stress and…

Spatial assessment of soluble solid contents on apple slices using hyperspectral imaging

USDA-ARS?s Scientific Manuscript database

A partial least squares regression (PLSR) model to map internal soluble solids content (SSC) of apples using visible/near-infrared (VNIR) hyperspectral imaging was developed. The reflectance spectra of sliced apples were extracted from hyperspectral absorbance images obtained in the 400e1000 nm rang...

The Use of Alternative Regression Methods in Social Sciences and the Comparison of Least Squares and M Estimation Methods in Terms of the Determination of Coefficient

ERIC Educational Resources Information Center

Coskuntuncel, Orkun

2013-01-01

The purpose of this study is two-fold; the first aim being to show the effect of outliers on the widely used least squares regression estimator in social sciences. The second aim is to compare the classical method of least squares with the robust M-estimator using the "determination of coefficient" (R[superscript 2]). For this purpose,…

Wind Tunnel Strain-Gage Balance Calibration Data Analysis Using a Weighted Least Squares Approach

NASA Technical Reports Server (NTRS)

Ulbrich, N.; Volden, T.

2017-01-01

A new approach is presented that uses a weighted least squares fit to analyze wind tunnel strain-gage balance calibration data. The weighted least squares fit is specifically designed to increase the influence of single-component loadings during the regression analysis. The weighted least squares fit also reduces the impact of calibration load schedule asymmetries on the predicted primary sensitivities of the balance gages. A weighting factor between zero and one is assigned to each calibration data point that depends on a simple count of its intentionally loaded load components or gages. The greater the number of a data point's intentionally loaded load components or gages is, the smaller its weighting factor becomes. The proposed approach is applicable to both the Iterative and Non-Iterative Methods that are used for the analysis of strain-gage balance calibration data in the aerospace testing community. The Iterative Method uses a reasonable estimate of the tare corrected load set as input for the determination of the weighting factors. The Non-Iterative Method, on the other hand, uses gage output differences relative to the natural zeros as input for the determination of the weighting factors. Machine calibration data of a six-component force balance is used to illustrate benefits of the proposed weighted least squares fit. In addition, a detailed derivation of the PRESS residuals associated with a weighted least squares fit is given in the appendices of the paper as this information could not be found in the literature. These PRESS residuals may be needed to evaluate the predictive capabilities of the final regression models that result from a weighted least squares fit of the balance calibration data.

Application of Partial Least Squares (PLS) Regression to Determine Landscape-Scale Aquatic Resource Vulnerability in the Ozark Mountains

EPA Science Inventory

Partial least squares (PLS) analysis offers a number of advantages over the more traditionally used regression analyses applied in landscape ecology to study the associations among constituents of surface water and landscapes. Common data problems in ecological studies include: s...

Healthcare Expenditures Associated with Depression Among Individuals with Osteoarthritis: Post-Regression Linear Decomposition Approach.

PubMed

Agarwal, Parul; Sambamoorthi, Usha

2015-12-01

Depression is common among individuals with osteoarthritis and leads to increased healthcare burden. The objective of this study was to examine excess total healthcare expenditures associated with depression among individuals with osteoarthritis in the US. Adults with self-reported osteoarthritis (n = 1881) were identified using data from the 2010 Medical Expenditure Panel Survey (MEPS). Among those with osteoarthritis, chi-square tests and ordinary least square regressions (OLS) were used to examine differences in healthcare expenditures between those with and without depression. Post-regression linear decomposition technique was used to estimate the relative contribution of different constructs of the Anderson's behavioral model, i.e., predisposing, enabling, need, personal healthcare practices, and external environment factors, to the excess expenditures associated with depression among individuals with osteoarthritis. All analysis accounted for the complex survey design of MEPS. Depression coexisted among 20.6 % of adults with osteoarthritis. The average total healthcare expenditures were $13,684 among adults with depression compared to $9284 among those without depression. Multivariable OLS regression revealed that adults with depression had 38.8 % higher healthcare expenditures (p < 0.001) compared to those without depression. Post-regression linear decomposition analysis indicated that 50 % of differences in expenditures among adults with and without depression can be explained by differences in need factors. Among individuals with coexisting osteoarthritis and depression, excess healthcare expenditures associated with depression were mainly due to comorbid anxiety, chronic conditions and poor health status. These expenditures may potentially be reduced by providing timely intervention for need factors or by providing care under a collaborative care model.

Identifying maternal and infant factors associated with newborn size in rural Bangladesh by partial least squares (PLS) regression analysis

PubMed Central

Rahman, Md. Jahanur; Shamim, Abu Ahmed; Klemm, Rolf D. W.; Labrique, Alain B.; Rashid, Mahbubur; Christian, Parul; West, Keith P.

2017-01-01

Birth weight, length and circumferences of the head, chest and arm are key measures of newborn size and health in developing countries. We assessed maternal socio-demographic factors associated with multiple measures of newborn size in a large rural population in Bangladesh using partial least squares (PLS) regression method. PLS regression, combining features from principal component analysis and multiple linear regression, is a multivariate technique with an ability to handle multicollinearity while simultaneously handling multiple dependent variables. We analyzed maternal and infant data from singletons (n = 14,506) born during a double-masked, cluster-randomized, placebo-controlled maternal vitamin A or β-carotene supplementation trial in rural northwest Bangladesh. PLS regression results identified numerous maternal factors (parity, age, early pregnancy MUAC, living standard index, years of education, number of antenatal care visits, preterm delivery and infant sex) significantly (p<0.001) associated with newborn size. Among them, preterm delivery had the largest negative influence on newborn size (Standardized β = -0.29 − -0.19; p<0.001). Scatter plots of the scores of first two PLS components also revealed an interaction between newborn sex and preterm delivery on birth size. PLS regression was found to be more parsimonious than both ordinary least squares regression and principal component regression. It also provided more stable estimates than the ordinary least squares regression and provided the effect measure of the covariates with greater accuracy as it accounts for the correlation among the covariates and outcomes. Therefore, PLS regression is recommended when either there are multiple outcome measurements in the same study, or the covariates are correlated, or both situations exist in a dataset. PMID:29261760

Identifying maternal and infant factors associated with newborn size in rural Bangladesh by partial least squares (PLS) regression analysis.

PubMed

Kabir, Alamgir; Rahman, Md Jahanur; Shamim, Abu Ahmed; Klemm, Rolf D W; Labrique, Alain B; Rashid, Mahbubur; Christian, Parul; West, Keith P

2017-01-01

Birth weight, length and circumferences of the head, chest and arm are key measures of newborn size and health in developing countries. We assessed maternal socio-demographic factors associated with multiple measures of newborn size in a large rural population in Bangladesh using partial least squares (PLS) regression method. PLS regression, combining features from principal component analysis and multiple linear regression, is a multivariate technique with an ability to handle multicollinearity while simultaneously handling multiple dependent variables. We analyzed maternal and infant data from singletons (n = 14,506) born during a double-masked, cluster-randomized, placebo-controlled maternal vitamin A or β-carotene supplementation trial in rural northwest Bangladesh. PLS regression results identified numerous maternal factors (parity, age, early pregnancy MUAC, living standard index, years of education, number of antenatal care visits, preterm delivery and infant sex) significantly (p<0.001) associated with newborn size. Among them, preterm delivery had the largest negative influence on newborn size (Standardized β = -0.29 - -0.19; p<0.001). Scatter plots of the scores of first two PLS components also revealed an interaction between newborn sex and preterm delivery on birth size. PLS regression was found to be more parsimonious than both ordinary least squares regression and principal component regression. It also provided more stable estimates than the ordinary least squares regression and provided the effect measure of the covariates with greater accuracy as it accounts for the correlation among the covariates and outcomes. Therefore, PLS regression is recommended when either there are multiple outcome measurements in the same study, or the covariates are correlated, or both situations exist in a dataset.

Guidelines and Procedures for Computing Time-Series Suspended-Sediment Concentrations and Loads from In-Stream Turbidity-Sensor and Streamflow Data

USGS Publications Warehouse

Rasmussen, Patrick P.; Gray, John R.; Glysson, G. Douglas; Ziegler, Andrew C.

2009-01-01

In-stream continuous turbidity and streamflow data, calibrated with measured suspended-sediment concentration data, can be used to compute a time series of suspended-sediment concentration and load at a stream site. Development of a simple linear (ordinary least squares) regression model for computing suspended-sediment concentrations from instantaneous turbidity data is the first step in the computation process. If the model standard percentage error (MSPE) of the simple linear regression model meets a minimum criterion, this model should be used to compute a time series of suspended-sediment concentrations. Otherwise, a multiple linear regression model using paired instantaneous turbidity and streamflow data is developed and compared to the simple regression model. If the inclusion of the streamflow variable proves to be statistically significant and the uncertainty associated with the multiple regression model results in an improvement over that for the simple linear model, the turbidity-streamflow multiple linear regression model should be used to compute a suspended-sediment concentration time series. The computed concentration time series is subsequently used with its paired streamflow time series to compute suspended-sediment loads by standard U.S. Geological Survey techniques. Once an acceptable regression model is developed, it can be used to compute suspended-sediment concentration beyond the period of record used in model development with proper ongoing collection and analysis of calibration samples. Regression models to compute suspended-sediment concentrations are generally site specific and should never be considered static, but they represent a set period in a continually dynamic system in which additional data will help verify any change in sediment load, type, and source.

Can Emotional and Behavioral Dysregulation in Youth Be Decoded from Functional Neuroimaging?

PubMed

Portugal, Liana C L; Rosa, Maria João; Rao, Anil; Bebko, Genna; Bertocci, Michele A; Hinze, Amanda K; Bonar, Lisa; Almeida, Jorge R C; Perlman, Susan B; Versace, Amelia; Schirda, Claudiu; Travis, Michael; Gill, Mary Kay; Demeter, Christine; Diwadkar, Vaibhav A; Ciuffetelli, Gary; Rodriguez, Eric; Forbes, Erika E; Sunshine, Jeffrey L; Holland, Scott K; Kowatch, Robert A; Birmaher, Boris; Axelson, David; Horwitz, Sarah M; Arnold, Eugene L; Fristad, Mary A; Youngstrom, Eric A; Findling, Robert L; Pereira, Mirtes; Oliveira, Leticia; Phillips, Mary L; Mourao-Miranda, Janaina

2016-01-01

High comorbidity among pediatric disorders characterized by behavioral and emotional dysregulation poses problems for diagnosis and treatment, and suggests that these disorders may be better conceptualized as dimensions of abnormal behaviors. Furthermore, identifying neuroimaging biomarkers related to dimensional measures of behavior may provide targets to guide individualized treatment. We aimed to use functional neuroimaging and pattern regression techniques to determine whether patterns of brain activity could accurately decode individual-level severity on a dimensional scale measuring behavioural and emotional dysregulation at two different time points. A sample of fifty-seven youth (mean age: 14.5 years; 32 males) was selected from a multi-site study of youth with parent-reported behavioral and emotional dysregulation. Participants performed a block-design reward paradigm during functional Magnetic Resonance Imaging (fMRI). Pattern regression analyses consisted of Relevance Vector Regression (RVR) and two cross-validation strategies implemented in the Pattern Recognition for Neuroimaging toolbox (PRoNTo). Medication was treated as a binary confounding variable. Decoded and actual clinical scores were compared using Pearson's correlation coefficient (r) and mean squared error (MSE) to evaluate the models. Permutation test was applied to estimate significance levels. Relevance Vector Regression identified patterns of neural activity associated with symptoms of behavioral and emotional dysregulation at the initial study screen and close to the fMRI scanning session. The correlation and the mean squared error between actual and decoded symptoms were significant at the initial study screen and close to the fMRI scanning session. However, after controlling for potential medication effects, results remained significant only for decoding symptoms at the initial study screen. Neural regions with the highest contribution to the pattern regression model included cerebellum, sensory-motor and fronto-limbic areas. The combination of pattern regression models and neuroimaging can help to determine the severity of behavioral and emotional dysregulation in youth at different time points.

Nonlinear Constitutive Modeling of Piezoelectric Ceramics

NASA Astrophysics Data System (ADS)

Xu, Jia; Li, Chao; Wang, Haibo; Zhu, Zhiwen

2017-12-01

Nonlinear constitutive modeling of piezoelectric ceramics is discussed in this paper. Van der Pol item is introduced to explain the simple hysteretic curve. Improved nonlinear difference items are used to interpret the hysteresis phenomena of piezoelectric ceramics. The fitting effect of the model on experimental data is proved by the partial least-square regression method. The results show that this method can describe the real curve well. The results of this paper are helpful to piezoelectric ceramics constitutive modeling.

Gender Gap in Mathematics and Physics in Chinese Middle Schools: A Case Study of A Beijing's District

ERIC Educational Resources Information Center

Li, Manli; Zhang, Yu; Wang, Yihan

2017-01-01

This study examines the gender gaps in mathematics and physics in Chinese middle schools. The data is from the Education Bureau management database which includes all middle school students who took high school entrance exam in a district of Beijing from 2006-2013. The ordinary least square model and quantile regression model are applied. This…

Gender Gap in the National College Entrance Exam Performance in China: A Case Study of a Typical Chinese Municipality

ERIC Educational Resources Information Center

Zhang, Yu; Tsang, Mun

2015-01-01

This is one of the first studies to investigate gender achievement gap in the National College Entrance Exam in a typical municipality in China, which is the crucial examination for the transition from high school to higher education in that country. Using ordinary least square model and quantile regression model, the study consistently finds that…

Relationship between Urbanization and Cancer Incidence in Iran Using Quantile Regression.

PubMed

Momenyan, Somayeh; Sadeghifar, Majid; Sarvi, Fatemeh; Khodadost, Mahmoud; Mosavi-Jarrahi, Alireza; Ghaffari, Mohammad Ebrahim; Sekhavati, Eghbal

2016-01-01

Quantile regression is an efficient method for predicting and estimating the relationship between explanatory variables and percentile points of the response distribution, particularly for extreme percentiles of the distribution. To study the relationship between urbanization and cancer morbidity, we here applied quantile regression. This cross-sectional study was conducted for 9 cancers in 345 cities in 2007 in Iran. Data were obtained from the Ministry of Health and Medical Education and the relationship between urbanization and cancer morbidity was investigated using quantile regression and least square regression. Fitting models were compared using AIC criteria. R (3.0.1) software and the Quantreg package were used for statistical analysis. With the quantile regression model all percentiles for breast, colorectal, prostate, lung and pancreas cancers demonstrated increasing incidence rate with urbanization. The maximum increase for breast cancer was in the 90th percentile (β=0.13, p-value<0.001), for colorectal cancer was in the 75th percentile (β=0.048, p-value<0.001), for prostate cancer the 95th percentile (β=0.55, p-value<0.001), for lung cancer was in 95th percentile (β=0.52, p-value=0.006), for pancreas cancer was in 10th percentile (β=0.011, p-value<0.001). For gastric, esophageal and skin cancers, with increasing urbanization, the incidence rate was decreased. The maximum decrease for gastric cancer was in the 90th percentile(β=0.003, p-value<0.001), for esophageal cancer the 95th (β=0.04, p-value=0.4) and for skin cancer also the 95th (β=0.145, p-value=0.071). The AIC showed that for upper percentiles, the fitting of quantile regression was better than least square regression. According to the results of this study, the significant impact of urbanization on cancer morbidity requirs more effort and planning by policymakers and administrators in order to reduce risk factors such as pollution in urban areas and ensure proper nutrition recommendations are made.

Modified multiblock partial least squares path modeling algorithm with backpropagation neural networks approach

NASA Astrophysics Data System (ADS)

Yuniarto, Budi; Kurniawan, Robert

2017-03-01

PLS Path Modeling (PLS-PM) is different from covariance based SEM, where PLS-PM use an approach based on variance or component, therefore, PLS-PM is also known as a component based SEM. Multiblock Partial Least Squares (MBPLS) is a method in PLS regression which can be used in PLS Path Modeling which known as Multiblock PLS Path Modeling (MBPLS-PM). This method uses an iterative procedure in its algorithm. This research aims to modify MBPLS-PM with Back Propagation Neural Network approach. The result is MBPLS-PM algorithm can be modified using the Back Propagation Neural Network approach to replace the iterative process in backward and forward step to get the matrix t and the matrix u in the algorithm. By modifying the MBPLS-PM algorithm using Back Propagation Neural Network approach, the model parameters obtained are relatively not significantly different compared to model parameters obtained by original MBPLS-PM algorithm.

Application of Visible and Near-Infrared Hyperspectral Imaging to Determine Soluble Protein Content in Oilseed Rape Leaves

PubMed Central

Zhang, Chu; Liu, Fei; Kong, Wenwen; He, Yong

2015-01-01

Visible and near-infrared hyperspectral imaging covering spectral range of 380–1030 nm as a rapid and non-destructive method was applied to estimate the soluble protein content of oilseed rape leaves. Average spectrum (500–900 nm) of the region of interest (ROI) of each sample was extracted, and four samples out of 128 samples were defined as outliers by Monte Carlo-partial least squares (MCPLS). Partial least squares (PLS) model using full spectra obtained dependable performance with the correlation coefficient (rp) of 0.9441, root mean square error of prediction (RMSEP) of 0.1658 mg/g and residual prediction deviation (RPD) of 2.98. The weighted regression coefficient (Bw), successive projections algorithm (SPA) and genetic algorithm-partial least squares (GAPLS) selected 18, 15, and 16 sensitive wavelengths, respectively. SPA-PLS model obtained the best performance with rp of 0.9554, RMSEP of 0.1538 mg/g and RPD of 3.25. Distribution of protein content within the rape leaves were visualized and mapped on the basis of the SPA-PLS model. The overall results indicated that hyperspectral imaging could be used to determine and visualize the soluble protein content of rape leaves. PMID:26184198

Perceptions and Efficacy of Flight Operational Quality Assurance (FOQA) Programs Among Small-scale Operators

DTIC Science & Technology

2012-01-01

regressive Integrated Moving Average ( ARIMA ) model for the data, eliminating the need to identify an appropriate model through trial and error alone...06 .11 13.67 16 .62 16 .14 .11 8.06 16 .95 * Based on the asymptotic chi-square approximation. 8 In general, ARIMA models address three...performance standards and measurement processes and a prevailing climate of organizational trust were important factors. Unfortunately, uneven

Forcing Regression through a Given Point Using Any Familiar Computational Routine.

DTIC Science & Technology

1983-03-01

a linear model , Y =a + OX + e ( Model I) then adopt the principle of least squares; and use sample data to estimate the unknown parameters, a and 8...has an expected value of zero indicates that the "average" response is considered linear . If c varies widely, Model I, though conceptually correct, may...relationship is linear from the maximum observed x to x - a, then Model II should be used. To pro- ceed with the customary evaluation of Model I would be

Methods for estimating the magnitude and frequency of peak streamflows for unregulated streams in Oklahoma

USGS Publications Warehouse

Lewis, Jason M.

2010-01-01

Peak-streamflow regression equations were determined for estimating flows with exceedance probabilities from 50 to 0.2 percent for the state of Oklahoma. These regression equations incorporate basin characteristics to estimate peak-streamflow magnitude and frequency throughout the state by use of a generalized least squares regression analysis. The most statistically significant independent variables required to estimate peak-streamflow magnitude and frequency for unregulated streams in Oklahoma are contributing drainage area, mean-annual precipitation, and main-channel slope. The regression equations are applicable for watershed basins with drainage areas less than 2,510 square miles that are not affected by regulation. The resulting regression equations had a standard model error ranging from 31 to 46 percent. Annual-maximum peak flows observed at 231 streamflow-gaging stations through water year 2008 were used for the regression analysis. Gage peak-streamflow estimates were used from previous work unless 2008 gaging-station data were available, in which new peak-streamflow estimates were calculated. The U.S. Geological Survey StreamStats web application was used to obtain the independent variables required for the peak-streamflow regression equations. Limitations on the use of the regression equations and the reliability of regression estimates for natural unregulated streams are described. Log-Pearson Type III analysis information, basin and climate characteristics, and the peak-streamflow frequency estimates for the 231 gaging stations in and near Oklahoma are listed. Methodologies are presented to estimate peak streamflows at ungaged sites by using estimates from gaging stations on unregulated streams. For ungaged sites on urban streams and streams regulated by small floodwater retarding structures, an adjustment of the statewide regression equations for natural unregulated streams can be used to estimate peak-streamflow magnitude and frequency.

GSTAR-SUR Modeling With Calendar Variations And Intervention To Forecast Outflow Of Currencies In Java Indonesia

NASA Astrophysics Data System (ADS)

Akbar, M. S.; Setiawan; Suhartono; Ruchjana, B. N.; Riyadi, M. A. A.

2018-03-01

Ordinary Least Squares (OLS) is general method to estimates Generalized Space Time Autoregressive (GSTAR) parameters. But in some cases, the residuals of GSTAR are correlated between location. If OLS is applied to this case, then the estimators are inefficient. Generalized Least Squares (GLS) is a method used in Seemingly Unrelated Regression (SUR) model. This method estimated parameters of some models with residuals between equations are correlated. Simulation study shows that GSTAR with GLS method for estimating parameters (GSTAR-SUR) is more efficient than GSTAR-OLS method. The purpose of this research is to apply GSTAR-SUR with calendar variation and intervention as exogenous variable (GSTARX-SUR) for forecast outflow of currency in Java, Indonesia. As a result, GSTARX-SUR provides better performance than GSTARX-OLS.

Comprehensive ripeness-index for prediction of ripening level in mangoes by multivariate modelling of ripening behaviour

NASA Astrophysics Data System (ADS)

Eyarkai Nambi, Vijayaram; Thangavel, Kuladaisamy; Manickavasagan, Annamalai; Shahir, Sultan

2017-01-01

Prediction of ripeness level in climacteric fruits is essential for post-harvest handling. An index capable of predicting ripening level with minimum inputs would be highly beneficial to the handlers, processors and researchers in fruit industry. A study was conducted with Indian mango cultivars to develop a ripeness index and associated model. Changes in physicochemical, colour and textural properties were measured throughout the ripening period and the period was classified into five stages (unripe, early ripe, partially ripe, ripe and over ripe). Multivariate regression techniques like partial least square regression, principal component regression and multi linear regression were compared and evaluated for its prediction. Multi linear regression model with 12 parameters was found more suitable in ripening prediction. Scientific variable reduction method was adopted to simplify the developed model. Better prediction was achieved with either 2 or 3 variables (total soluble solids, colour and acidity). Cross validation was done to increase the robustness and it was found that proposed ripening index was more effective in prediction of ripening stages. Three-variable model would be suitable for commercial applications where reasonable accuracies are sufficient. However, 12-variable model can be used to obtain more precise results in research and development applications.

The comparison between several robust ridge regression estimators in the presence of multicollinearity and multiple outliers

NASA Astrophysics Data System (ADS)

Zahari, Siti Meriam; Ramli, Norazan Mohamed; Moktar, Balkiah; Zainol, Mohammad Said

2014-09-01

In the presence of multicollinearity and multiple outliers, statistical inference of linear regression model using ordinary least squares (OLS) estimators would be severely affected and produces misleading results. To overcome this, many approaches have been investigated. These include robust methods which were reported to be less sensitive to the presence of outliers. In addition, ridge regression technique was employed to tackle multicollinearity problem. In order to mitigate both problems, a combination of ridge regression and robust methods was discussed in this study. The superiority of this approach was examined when simultaneous presence of multicollinearity and multiple outliers occurred in multiple linear regression. This study aimed to look at the performance of several well-known robust estimators; M, MM, RIDGE and robust ridge regression estimators, namely Weighted Ridge M-estimator (WRM), Weighted Ridge MM (WRMM), Ridge MM (RMM), in such a situation. Results of the study showed that in the presence of simultaneous multicollinearity and multiple outliers (in both x and y-direction), the RMM and RIDGE are more or less similar in terms of superiority over the other estimators, regardless of the number of observation, level of collinearity and percentage of outliers used. However, when outliers occurred in only single direction (y-direction), the WRMM estimator is the most superior among the robust ridge regression estimators, by producing the least variance. In conclusion, the robust ridge regression is the best alternative as compared to robust and conventional least squares estimators when dealing with simultaneous presence of multicollinearity and outliers.

Confidence Intervals for Squared Semipartial Correlation Coefficients: The Effect of Nonnormality

ERIC Educational Resources Information Center

Algina, James; Keselman, H. J.; Penfield, Randall D.

2010-01-01

The increase in the squared multiple correlation coefficient ([delta]R[superscript 2]) associated with a variable in a regression equation is a commonly used measure of importance in regression analysis. Algina, Keselman, and Penfield found that intervals based on asymptotic principles were typically very inaccurate, even though the sample size…

A Comparison of Mean Phase Difference and Generalized Least Squares for Analyzing Single-Case Data

ERIC Educational Resources Information Center

Manolov, Rumen; Solanas, Antonio

2013-01-01

The present study focuses on single-case data analysis specifically on two procedures for quantifying differences between baseline and treatment measurements. The first technique tested is based on generalized least square regression analysis and is compared to a proposed non-regression technique, which allows obtaining similar information. The…

Improving near-infrared prediction model robustness with support vector machine regression: a pharmaceutical tablet assay example.

PubMed

Igne, Benoît; Drennen, James K; Anderson, Carl A

2014-01-01

Changes in raw materials and process wear and tear can have significant effects on the prediction error of near-infrared calibration models. When the variability that is present during routine manufacturing is not included in the calibration, test, and validation sets, the long-term performance and robustness of the model will be limited. Nonlinearity is a major source of interference. In near-infrared spectroscopy, nonlinearity can arise from light path-length differences that can come from differences in particle size or density. The usefulness of support vector machine (SVM) regression to handle nonlinearity and improve the robustness of calibration models in scenarios where the calibration set did not include all the variability present in test was evaluated. Compared to partial least squares (PLS) regression, SVM regression was less affected by physical (particle size) and chemical (moisture) differences. The linearity of the SVM predicted values was also improved. Nevertheless, although visualization and interpretation tools have been developed to enhance the usability of SVM-based methods, work is yet to be done to provide chemometricians in the pharmaceutical industry with a regression method that can supplement PLS-based methods.

Use of AMMI and linear regression models to analyze genotype-environment interaction in durum wheat.

PubMed

Nachit, M M; Nachit, G; Ketata, H; Gauch, H G; Zobel, R W

1992-03-01

The joint durum wheat (Triticum turgidum L var 'durum') breeding program of the International Maize and Wheat Improvement Center (CIMMYT) and the International Center for Agricultural Research in the Dry Areas (ICARDA) for the Mediterranean region employs extensive multilocation testing. Multilocation testing produces significant genotype-environment (GE) interaction that reduces the accuracy for estimating yield and selecting appropriate germ plasm. The sum of squares (SS) of GE interaction was partitioned by linear regression techniques into joint, genotypic, and environmental regressions, and by Additive Main effects and the Multiplicative Interactions (AMMI) model into five significant Interaction Principal Component Axes (IPCA). The AMMI model was more effective in partitioning the interaction SS than the linear regression technique. The SS contained in the AMMI model was 6 times higher than the SS for all three regressions. Postdictive assessment recommended the use of the first five IPCA axes, while predictive assessment AMMI1 (main effects plus IPCA1). After elimination of random variation, AMMI1 estimates for genotypic yields within sites were more precise than unadjusted means. This increased precision was equivalent to increasing the number of replications by a factor of 3.7.

Weather variability and the incidence of cryptosporidiosis: comparison of time series poisson regression and SARIMA models.

PubMed

Hu, Wenbiao; Tong, Shilu; Mengersen, Kerrie; Connell, Des

2007-09-01

Few studies have examined the relationship between weather variables and cryptosporidiosis in Australia. This paper examines the potential impact of weather variability on the transmission of cryptosporidiosis and explores the possibility of developing an empirical forecast system. Data on weather variables, notified cryptosporidiosis cases, and population size in Brisbane were supplied by the Australian Bureau of Meteorology, Queensland Department of Health, and Australian Bureau of Statistics for the period of January 1, 1996-December 31, 2004, respectively. Time series Poisson regression and seasonal auto-regression integrated moving average (SARIMA) models were performed to examine the potential impact of weather variability on the transmission of cryptosporidiosis. Both the time series Poisson regression and SARIMA models show that seasonal and monthly maximum temperature at a prior moving average of 1 and 3 months were significantly associated with cryptosporidiosis disease. It suggests that there may be 50 more cases a year for an increase of 1 degrees C maximum temperature on average in Brisbane. Model assessments indicated that the SARIMA model had better predictive ability than the Poisson regression model (SARIMA: root mean square error (RMSE): 0.40, Akaike information criterion (AIC): -12.53; Poisson regression: RMSE: 0.54, AIC: -2.84). Furthermore, the analysis of residuals shows that the time series Poisson regression appeared to violate a modeling assumption, in that residual autocorrelation persisted. The results of this study suggest that weather variability (particularly maximum temperature) may have played a significant role in the transmission of cryptosporidiosis. A SARIMA model may be a better predictive model than a Poisson regression model in the assessment of the relationship between weather variability and the incidence of cryptosporidiosis.

Local-scale spatial modelling for interpolating climatic temperature variables to predict agricultural plant suitability

NASA Astrophysics Data System (ADS)

Webb, Mathew A.; Hall, Andrew; Kidd, Darren; Minansy, Budiman

2016-05-01

Assessment of local spatial climatic variability is important in the planning of planting locations for horticultural crops. This study investigated three regression-based calibration methods (i.e. traditional versus two optimized methods) to relate short-term 12-month data series from 170 temperature loggers and 4 weather station sites with data series from nearby long-term Australian Bureau of Meteorology climate stations. The techniques trialled to interpolate climatic temperature variables, such as frost risk, growing degree days (GDDs) and chill hours, were regression kriging (RK), regression trees (RTs) and random forests (RFs). All three calibration methods produced accurate results, with the RK-based calibration method delivering the most accurate validation measures: coefficients of determination ( R 2) of 0.92, 0.97 and 0.95 and root-mean-square errors of 1.30, 0.80 and 1.31 °C, for daily minimum, daily maximum and hourly temperatures, respectively. Compared with the traditional method of calibration using direct linear regression between short-term and long-term stations, the RK-based calibration method improved R 2 and reduced root-mean-square error (RMSE) by at least 5 % and 0.47 °C for daily minimum temperature, 1 % and 0.23 °C for daily maximum temperature and 3 % and 0.33 °C for hourly temperature. Spatial modelling indicated insignificant differences between the interpolation methods, with the RK technique tending to be the slightly better method due to the high degree of spatial autocorrelation between logger sites.

Parametric output-only identification of time-varying structures using a kernel recursive extended least squares TARMA approach

NASA Astrophysics Data System (ADS)

Ma, Zhi-Sai; Liu, Li; Zhou, Si-Da; Yu, Lei; Naets, Frank; Heylen, Ward; Desmet, Wim

2018-01-01

The problem of parametric output-only identification of time-varying structures in a recursive manner is considered. A kernelized time-dependent autoregressive moving average (TARMA) model is proposed by expanding the time-varying model parameters onto the basis set of kernel functions in a reproducing kernel Hilbert space. An exponentially weighted kernel recursive extended least squares TARMA identification scheme is proposed, and a sliding-window technique is subsequently applied to fix the computational complexity for each consecutive update, allowing the method to operate online in time-varying environments. The proposed sliding-window exponentially weighted kernel recursive extended least squares TARMA method is employed for the identification of a laboratory time-varying structure consisting of a simply supported beam and a moving mass sliding on it. The proposed method is comparatively assessed against an existing recursive pseudo-linear regression TARMA method via Monte Carlo experiments and shown to be capable of accurately tracking the time-varying dynamics. Furthermore, the comparisons demonstrate the superior achievable accuracy, lower computational complexity and enhanced online identification capability of the proposed kernel recursive extended least squares TARMA approach.

Estimating the magnitude of annual peak discharges with recurrence intervals between 1.1 and 3.0 years for rural, unregulated streams in West Virginia

USGS Publications Warehouse

Wiley, Jeffrey B.; Atkins, John T.; Newell, Dawn A.

2002-01-01

Multiple and simple least-squares regression models for the log10-transformed 1.5- and 2-year recurrence intervals of peak discharges with independent variables describing the basin characteristics (log10-transformed and untransformed) for 236 streamflow-gaging stations were evaluated, and the regression residuals were plotted as areal distributions that defined three regions in West Virginia designated as East, North, and South. Regional equations for the 1.1-, 1.2-, 1.3-, 1.4-, 1.5-, 1.6-, 1.7-, 1.8-, 1.9-, 2.0-, 2.5-, and 3-year recurrence intervals of peak discharges were determined by generalized least-squares regression. Log10-transformed drainage area was the most significant independent variable for all regions. Equations developed in this study are applicable only to rural, unregulated streams within the boundaries of West Virginia. The accuracies of estimating equations are quantified by measuring the average prediction error (from 27.4 to 52.4 percent) and equivalent years of record (from 1.1 to 3.4 years).

Rapid quantification of casein in skim milk using Fourier transform infrared spectroscopy, enzymatic perturbation, and multiway partial least squares regression: Monitoring chymosin at work.

PubMed

Baum, A; Hansen, P W; Nørgaard, L; Sørensen, John; Mikkelsen, J D

2016-08-01

In this study, we introduce enzymatic perturbation combined with Fourier transform infrared (FTIR) spectroscopy as a concept for quantifying casein in subcritical heated skim milk using chemometric multiway analysis. Chymosin is a protease that cleaves specifically caseins. As a result of hydrolysis, all casein proteins clot to form a creamy precipitate, and whey proteins remain in the supernatant. We monitored the cheese-clotting reaction in real time using FTIR and analyzed the resulting evolution profiles to establish calibration models using parallel factor analysis and multiway partial least squares regression. Because we observed casein-specific kinetic changes, the retrieved models were independent of the chemical background matrix and were therefore robust against possible covariance effects. We tested the robustness of the models by spiking the milk solutions with whey, calcium, and cream. This method can be used at different stages in the dairy production chain to ensure the quality of the delivered milk. In particular, the cheese-making industry can benefit from such methods to optimize production control. Copyright © 2016 American Dairy Science Association. Published by Elsevier Inc. All rights reserved.

Techniques for estimating magnitude and frequency of peak flows for Pennsylvania streams

USGS Publications Warehouse

Stuckey, Marla H.; Reed, Lloyd A.

2000-01-01

Regression equations for estimating the magnitude and frequency of floods on ungaged streams in Pennsylvania with drainage areas less that 2,000 square miles were developed on the basis of peak-flow data collected at 313 streamflow-gaging stations. All streamflow-gaging stations used in the development of the equations had 10 or more years of record and include active and discontinued continuous-record and crest-stage partial-record streamflow-gaging stations. Regional regression equations were developed for flood flows expected every 10, 25, 50, 100, and 500 years by the use of a weighted multiple linear regression model.The State was divided into two regions. The largest region, Region A, encompasses about 78 percent of Pennsylvania. The smaller region, Region B, includes only the northwestern part of the State. Basin characteristics used in the regression equations for Region A are drainage area, percentage of forest cover, percentage of urban development, percentage of basin underlain by carbonate bedrock, and percentage of basin controlled by lakes, swamps, and reservoirs. Basin characteristics used in the regression equations for Region B are drainage area and percentage of basin controlled by lakes, swamps, and reservoirs. The coefficient of determination (R2) values for the five flood-frequency equations for Region A range from 0.93 to 0.82, and for Region B, the range is from 0.96 to 0.89.While the regression equations can be used to predict the magnitude and frequency of peak flows for most streams in the State, they should not be used for streams with drainage areas greater than 2,000 square miles or less than 1.5 square miles, for streams that drain extensively mined areas, or for stream reaches immediately below flood-control reservoirs. In addition, the equations presented for Region B should not be used if the stream drains a basin with more than 5 percent urban development.

Classification and quantitation of milk powder by near-infrared spectroscopy and mutual information-based variable selection and partial least squares

NASA Astrophysics Data System (ADS)

Chen, Hui; Tan, Chao; Lin, Zan; Wu, Tong

2018-01-01

Milk is among the most popular nutrient source worldwide, which is of great interest due to its beneficial medicinal properties. The feasibility of the classification of milk powder samples with respect to their brands and the determination of protein concentration is investigated by NIR spectroscopy along with chemometrics. Two datasets were prepared for experiment. One contains 179 samples of four brands for classification and the other contains 30 samples for quantitative analysis. Principal component analysis (PCA) was used for exploratory analysis. Based on an effective model-independent variable selection method, i.e., minimal-redundancy maximal-relevance (MRMR), only 18 variables were selected to construct a partial least-square discriminant analysis (PLS-DA) model. On the test set, the PLS-DA model based on the selected variable set was compared with the full-spectrum PLS-DA model, both of which achieved 100% accuracy. In quantitative analysis, the partial least-square regression (PLSR) model constructed by the selected subset of 260 variables outperforms significantly the full-spectrum model. It seems that the combination of NIR spectroscopy, MRMR and PLS-DA or PLSR is a powerful tool for classifying different brands of milk and determining the protein content.

Development of a hybrid proximal sensing method for rapid identification of petroleum contaminated soils.

PubMed

Chakraborty, Somsubhra; Weindorf, David C; Li, Bin; Ali Aldabaa, Abdalsamad Abdalsatar; Ghosh, Rakesh Kumar; Paul, Sathi; Nasim Ali, Md

2015-05-01

Using 108 petroleum contaminated soil samples, this pilot study proposed a new analytical approach of combining visible near-infrared diffuse reflectance spectroscopy (VisNIR DRS) and portable X-ray fluorescence spectrometry (PXRF) for rapid and improved quantification of soil petroleum contamination. Results indicated that an advanced fused model where VisNIR DRS spectra-based penalized spline regression (PSR) was used to predict total petroleum hydrocarbon followed by PXRF elemental data-based random forest regression was used to model the PSR residuals, it outperformed (R(2)=0.78, residual prediction deviation (RPD)=2.19) all other models tested, even producing better generalization than using VisNIR DRS alone (RPD's of 1.64, 1.86, and 1.96 for random forest, penalized spline regression, and partial least squares regression, respectively). Additionally, unsupervised principal component analysis using the PXRF+VisNIR DRS system qualitatively separated contaminated soils from control samples. Fusion of PXRF elemental data and VisNIR derivative spectra produced an optimized model for total petroleum hydrocarbon quantification in soils. Copyright © 2015 Elsevier B.V. All rights reserved.

On-line Model Structure Selection for Estimation of Plasma Boundary in a Tokamak

NASA Astrophysics Data System (ADS)

Škvára, Vít; Šmídl, Václav; Urban, Jakub

2015-11-01

Control of the plasma field in the tokamak requires reliable estimation of the plasma boundary. The plasma boundary is given by a complex mathematical model and the only available measurements are responses of induction coils around the plasma. For the purpose of boundary estimation the model can be reduced to simple linear regression with potentially infinitely many elements. The number of elements must be selected manually and this choice significantly influences the resulting shape. In this paper, we investigate the use of formal model structure estimation techniques for the problem. Specifically, we formulate a sparse least squares estimator using the automatic relevance principle. The resulting algorithm is a repetitive evaluation of the least squares problem which could be computed in real time. Performance of the resulting algorithm is illustrated on simulated data and evaluated with respect to a more detailed and computationally costly model FREEBIE.

Determination of volatile compounds in wine by gas chromatography-flame ionization detection: comparison between the U.S. Environmental Protection Agency 3sigma approach and Hubaux-Vos calculation of detection limits using ordinary and bivariate least squares.

PubMed

Caruso, Rosario; Scordino, Monica; Traulo, Pasqualino; Gagliano, Giacomo

2012-01-01

A capillary GC-flame ionization detection (FID) method to determine volatile compounds (ethyl acetate, 1,1-diethoxyethane, methyl alcohol, 1-propanol, 2-methyl-1-propanol, 2-methyl-1-butanol, 3-methyl-1-butanol, 1-butanol, and 2-butanol) in wine was investigated in terms of calculation of detection limits and calibration method. The main objectives were: (1) calculation of regression coefficient parameters by ordinary least-squares (OLS) and bivariate least-squares (BLS) regression models, taking into account errors in both axes; (2) estimation of linear dynamic range (LDR) according to International Conference on Harmonization recommendations; (3) performance evaluation of a method by using three different internal standards (ISs) such as acetonitrile, acetone, and 1-pentanol; (4) evaluation of LODs according to the U.S. Environmental Protection Agency (EPA) 3sigma approach and the Hubaux-Vos (H-V) method; (5) application of H-V theory to a gas chromatographic analytical method and to a food matrix; and (6) accuracy assessment of the method relative to methyl alcohol content through a Unione Italiana Vini (UIV) interlaboratory proficiency test. Calibration curves calculated via BLS and OLS show similar slopes, while intercepts are closer to zero in the first case, independent of the chosen IS. The studied ISs show a substantially equivalent behavior, even though the IS closer to the analyte retention time seems to be more appropriate in terms of LDR and LOD. Results indicate an underestimation of LODs using the EPA 3sigma approach instead of the more realistic H-V method, both with OLS and BLS regression models. Methanol contents compared with UIV average values indicate recovery between 90 and 110%.

Fearful and Distracted in School: Predicting Bullying among Youths

ERIC Educational Resources Information Center

Brewer, Steven Lawrence, Jr.; Meckley-Brewer, Hannah; Stinson, Philip M.

2017-01-01

Bullying and aggression in schools can have a traumatic and lasting effect on the well-being of children and youths. Using data from the 2013 National Crime Victimization Survey's School Crime Supplement, this study uses a chi-square automatic interaction detection (CHAID) decision tree and logistic regression models to identify factors that…

Bootstrapping Cox’s Regression Model.

DTIC Science & Technology

1985-11-01

crucial points a multivariate martingale central limit theorem. Involved in this is a p x p covariance matrix Z with elements T j2= f {2(s8 ) - s(l)( s ,8o...1980). The statistical analaysis of failure time data. Wiley, New York. Meyer, P.-A. (1971). Square integrable martingales, a survey. Lecture Notes

[Integral assessment of learning subjects difficulties].

PubMed

Grebniak, N P; Shchudro, S A

2010-01-01

The integral criterion for subject difficulties in senior classes is substantiated in terms of progress in studies, variation coefficient, and subjective and expert appraisals of the difficulty of subjects. The compiled regression models adequately determine the difficulty of academic subjects. According to the root-mean-square deviation, all subjects were found to have 3 degrees of difficulty.

Academic Performance of Community College Transfers: Psychological, Sociodemographic, and Educational Correlates

ERIC Educational Resources Information Center

Wang, Xueli

2012-01-01

This study focuses on the academic performance of community college transfer students at four-year institutions. It uses a nationally representative sample from the National Education Longitudinal Study of 1988 (NELS: 88/2000) and the Postsecondary Education Transcript Study (PETS). Results from an Ordinary Least Squares regression model suggest…

What Is the Relationship between Teacher Quality and Student Achievement? An Exploratory Study

ERIC Educational Resources Information Center

Stronge, James H.; Ward, Thomas J.; Tucker, Pamela D.; Hindman, Jennifer L.

2007-01-01

The major purpose of the study was to examine what constitutes effective teaching as defined by measured increases in student learning with a focus on the instructional behaviors and practices. Ordinary least squares (OLS) regression analyses and hierarchical linear modeling (HLM) were used to identify teacher effectiveness levels while…

Normalization Ridge Regression in Practice II: The Estimation of Multiple Feedback Linkages.

ERIC Educational Resources Information Center

Bulcock, J. W.

The use of the two-stage least squares (2 SLS) procedure for estimating nonrecursive social science models is often impractical when multiple feedback linkages are required. This is because 2 SLS is extremely sensitive to multicollinearity. The standard statistical solution to the multicollinearity problem is a biased, variance reduced procedure…

The Relationship of Institutional Tuition Discounts with Enrollment at Private, Not-for-Profit Institutions

ERIC Educational Resources Information Center

Lassila, Nathan E.

2010-01-01

Empirical studies exploring the impact of student aid on postsecondary enrollment often stop short of the specific examination of institutional tuition discounting. This research uses separate empirical ordinary least squares (OLS) regression models to examine three questions using public choice theory, positing that enrollment decisions may be…

Orthogonalizing EM: A design-based least squares algorithm

PubMed Central

Xiong, Shifeng; Dai, Bin; Huling, Jared; Qian, Peter Z. G.

2016-01-01

We introduce an efficient iterative algorithm, intended for various least squares problems, based on a design of experiments perspective. The algorithm, called orthogonalizing EM (OEM), works for ordinary least squares and can be easily extended to penalized least squares. The main idea of the procedure is to orthogonalize a design matrix by adding new rows and then solve the original problem by embedding the augmented design in a missing data framework. We establish several attractive theoretical properties concerning OEM. For the ordinary least squares with a singular regression matrix, an OEM sequence converges to the Moore-Penrose generalized inverse-based least squares estimator. For ordinary and penalized least squares with various penalties, it converges to a point having grouping coherence for fully aliased regression matrices. Convergence and the convergence rate of the algorithm are examined. Finally, we demonstrate that OEM is highly efficient for large-scale least squares and penalized least squares problems, and is considerably faster than competing methods when n is much larger than p. Supplementary materials for this article are available online. PMID:27499558

Squared exponential covariance function for prediction of hydrocarbon in seabed logging application

NASA Astrophysics Data System (ADS)

Mukhtar, Siti Mariam; Daud, Hanita; Dass, Sarat Chandra

2016-11-01

Seabed Logging technology (SBL) has progressively emerged as one of the demanding technologies in Exploration and Production (E&P) industry. Hydrocarbon prediction in deep water areas is crucial task for a driller in any oil and gas company as drilling cost is very expensive. Simulation data generated by Computer Software Technology (CST) is used to predict the presence of hydrocarbon where the models replicate real SBL environment. These models indicate that the hydrocarbon filled reservoirs are more resistive than surrounding water filled sediments. Then, as hydrocarbon depth is increased, it is more challenging to differentiate data with and without hydrocarbon. MATLAB is used for data extractions for curve fitting process using Gaussian process (GP). GP can be classified into regression and classification problems, where this work only focuses on Gaussian process regression (GPR) problem. Most popular choice to supervise GPR is squared exponential (SE), as it provides stability and probabilistic prediction in huge amounts of data. Hence, SE is used to predict the presence or absence of hydrocarbon in the reservoir from the data generated.

A rapid method for detection of fumonisins B1 and B2 in corn meal using Fourier transform near infrared (FT-NIR) spectroscopy implemented with integrating sphere.

PubMed

Gaspardo, B; Del Zotto, S; Torelli, E; Cividino, S R; Firrao, G; Della Riccia, G; Stefanon, B

2012-12-01

Fourier transform near infrared (FT-NIR) spectroscopy is an analytical procedure generally used to detect organic compounds in food. In this work the ability to predict fumonisin B(1)+B(2) contents in corn meal using an FT-NIR spectrophotometer, equipped with an integration sphere, was assessed. A total of 143 corn meal samples were collected in Friuli Venezia Giulia Region (Italy) and used to define a 15 principal components regression model, applying partial least square regression algorithm with full cross validation as internal validation. External validation was performed to 25 unknown samples. Coefficients of correlation, root mean square error and standard error of calibration were 0.964, 0.630 and 0.632, respectively and the external validation confirmed a fair potential of the model in predicting FB(1)+FB(2) concentration. Results suggest that FT-NIR analysis is a suitable method to detect FB(1)+FB(2) in corn meal and to discriminate safe meals from those contaminated. Copyright © 2012 Elsevier Ltd. All rights reserved.

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.

Comparison of Regression Methods to Compute Atmospheric Pressure and Earth Tidal Coefficients in Water Level Associated with Wenchuan Earthquake of 12 May 2008

NASA Astrophysics Data System (ADS)

He, Anhua; Singh, Ramesh P.; Sun, Zhaohua; Ye, Qing; Zhao, Gang

2016-07-01

The earth tide, atmospheric pressure, precipitation and earthquake fluctuations, especially earthquake greatly impacts water well levels, thus anomalous co-seismic changes in ground water levels have been observed. In this paper, we have used four different models, simple linear regression (SLR), multiple linear regression (MLR), principal component analysis (PCA) and partial least squares (PLS) to compute the atmospheric pressure and earth tidal effects on water level. Furthermore, we have used the Akaike information criterion (AIC) to study the performance of various models. Based on the lowest AIC and sum of squares for error values, the best estimate of the effects of atmospheric pressure and earth tide on water level is found using the MLR model. However, MLR model does not provide multicollinearity between inputs, as a result the atmospheric pressure and earth tidal response coefficients fail to reflect the mechanisms associated with the groundwater level fluctuations. On the premise of solving serious multicollinearity of inputs, PLS model shows the minimum AIC value. The atmospheric pressure and earth tidal response coefficients show close response with the observation using PLS model. The atmospheric pressure and the earth tidal response coefficients are found to be sensitive to the stress-strain state using the observed data for the period 1 April-8 June 2008 of Chuan 03# well. The transient enhancement of porosity of rock mass around Chuan 03# well associated with the Wenchuan earthquake (Mw = 7.9 of 12 May 2008) that has taken its original pre-seismic level after 13 days indicates that the co-seismic sharp rise of water well could be induced by static stress change, rather than development of new fractures.

A surrogate model for thermal characteristics of stratospheric airship

NASA Astrophysics Data System (ADS)

Zhao, Da; Liu, Dongxu; Zhu, Ming

2018-06-01

A simple and accurate surrogate model is extremely needed to reduce the analysis complexity of thermal characteristics for a stratospheric airship. In this paper, a surrogate model based on the Least Squares Support Vector Regression (LSSVR) is proposed. The Gravitational Search Algorithm (GSA) is used to optimize hyper parameters. A novel framework consisting of a preprocessing classifier and two regression models is designed to train the surrogate model. Various temperature datasets of the airship envelope and the internal gas are obtained by a three-dimensional transient model for thermal characteristics. Using these thermal datasets, two-factor and multi-factor surrogate models are trained and several comparison simulations are conducted. Results illustrate that the surrogate models based on LSSVR-GSA have good fitting and generalization abilities. The pre-treated classification strategy proposed in this paper plays a significant role in improving the accuracy of the surrogate model.

4D-Fingerprint Categorical QSAR Models for Skin Sensitization Based on Classification Local Lymph Node Assay Measures

PubMed Central

Li, Yi; Tseng, Yufeng J.; Pan, Dahua; Liu, Jianzhong; Kern, Petra S.; Gerberick, G. Frank; Hopfinger, Anton J.

2008-01-01

Currently, the only validated methods to identify skin sensitization effects are in vivo models, such as the Local Lymph Node Assay (LLNA) and guinea pig studies. There is a tremendous need, in particular due to novel legislation, to develop animal alternatives, eg. Quantitative Structure-Activity Relationship (QSAR) models. Here, QSAR models for skin sensitization using LLNA data have been constructed. The descriptors used to generate these models are derived from the 4D-molecular similarity paradigm and are referred to as universal 4D-fingerprints. A training set of 132 structurally diverse compounds and a test set of 15 structurally diverse compounds were used in this study. The statistical methodologies used to build the models are logistic regression (LR), and partial least square coupled logistic regression (PLS-LR), which prove to be effective tools for studying skin sensitization measures expressed in the two categorical terms of sensitizer and non-sensitizer. QSAR models with low values of the Hosmer-Lemeshow goodness-of-fit statistic, χHL2, are significant and predictive. For the training set, the cross-validated prediction accuracy of the logistic regression models ranges from 77.3% to 78.0%, while that of PLS-logistic regression models ranges from 87.1% to 89.4%. For the test set, the prediction accuracy of logistic regression models ranges from 80.0%-86.7%, while that of PLS-logistic regression models ranges from 73.3%-80.0%. The QSAR models are made up of 4D-fingerprints related to aromatic atoms, hydrogen bond acceptors and negatively partially charged atoms. PMID:17226934

Determining a Prony Series for a Viscoelastic Material From Time Varying Strain Data

NASA Technical Reports Server (NTRS)

Tzikang, Chen

2000-01-01

In this study a method of determining the coefficients in a Prony series representation of a viscoelastic modulus from rate dependent data is presented. Load versus time test data for a sequence of different rate loading segments is least-squares fitted to a Prony series hereditary integral model of the material tested. A nonlinear least squares regression algorithm is employed. The measured data includes ramp loading, relaxation, and unloading stress-strain data. The resulting Prony series which captures strain rate loading and unloading effects, produces an excellent fit to the complex loading sequence.

Quick method (FT-NIR) for the determination of oil and major fatty acids content in whole achenes of milk thistle (Silybum marianum (L.) Gaertn.).

PubMed

Koláčková, Pavla; Růžičková, Gabriela; Gregor, Tomáš; Šišperová, Eliška

2015-08-30

Calibration models for the Fourier transform-near infrared (FT-NIR) instrument were developed for quick and non-destructive determination of oil and fatty acids in whole achenes of milk thistle. Samples with a range of oil and fatty acid levels were collected and their transmittance spectra were obtained by the FT-NIR instrument. Based on these spectra and data gained by the means of the reference method - Soxhlet extraction and gas chromatography (GC) - calibration models were created by means of partial least square (PLS) regression analysis. Precision and accuracy of the calibration models was verified via the cross-validation of validation samples whose spectra were not part of the calibration model and also according to the root mean square error of prediction (RMSEP), root mean square error of calibration (RMSEC), root mean square error of cross-validation (RMSECV) and the validation coefficient of determination (R(2) ). R(2) for whole seeds were 0.96, 0.96, 0.83 and 0.67 and the RMSEP values were 0.76, 1.68, 1.24, 0.54 for oil, linoleic (C18:2), oleic (C18:1) and palmitic (C16:0) acids, respectively. The calibration models are appropriate for the non-destructive determination of oil and fatty acids levels in whole seeds of milk thistle. © 2014 Society of Chemical Industry.

[Application of artificial neural networks on the prediction of surface ozone concentrations].

PubMed

Shen, Lu-Lu; Wang, Yu-Xuan; Duan, Lei

2011-08-01

Ozone is an important secondary air pollutant in the lower atmosphere. In order to predict the hourly maximum ozone one day in advance based on the meteorological variables for the Wanqingsha site in Guangzhou, Guangdong province, a neural network model (Multi-Layer Perceptron) and a multiple linear regression model were used and compared. Model inputs are meteorological parameters (wind speed, wind direction, air temperature, relative humidity, barometric pressure and solar radiation) of the next day and hourly maximum ozone concentration of the previous day. The OBS (optimal brain surgeon) was adopted to prune the neutral work, to reduce its complexity and to improve its generalization ability. We find that the pruned neural network has the capacity to predict the peak ozone, with an agreement index of 92.3%, the root mean square error of 0.0428 mg/m3, the R-square of 0.737 and the success index of threshold exceedance 77.0% (the threshold O3 mixing ratio of 0.20 mg/m3). When the neural classifier was added to the neural network model, the success index of threshold exceedance increased to 83.6%. Through comparison of the performance indices between the multiple linear regression model and the neural network model, we conclud that that neural network is a better choice to predict peak ozone from meteorological forecast, which may be applied to practical prediction of ozone concentration.

A study of machine learning regression methods for major elemental analysis of rocks using laser-induced breakdown spectroscopy

NASA Astrophysics Data System (ADS)

Boucher, Thomas F.; Ozanne, Marie V.; Carmosino, Marco L.; Dyar, M. Darby; Mahadevan, Sridhar; Breves, Elly A.; Lepore, Kate H.; Clegg, Samuel M.

2015-05-01

The ChemCam instrument on the Mars Curiosity rover is generating thousands of LIBS spectra and bringing interest in this technique to public attention. The key to interpreting Mars or any other types of LIBS data are calibrations that relate laboratory standards to unknowns examined in other settings and enable predictions of chemical composition. Here, LIBS spectral data are analyzed using linear regression methods including partial least squares (PLS-1 and PLS-2), principal component regression (PCR), least absolute shrinkage and selection operator (lasso), elastic net, and linear support vector regression (SVR-Lin). These were compared against results from nonlinear regression methods including kernel principal component regression (K-PCR), polynomial kernel support vector regression (SVR-Py) and k-nearest neighbor (kNN) regression to discern the most effective models for interpreting chemical abundances from LIBS spectra of geological samples. The results were evaluated for 100 samples analyzed with 50 laser pulses at each of five locations averaged together. Wilcoxon signed-rank tests were employed to evaluate the statistical significance of differences among the nine models using their predicted residual sum of squares (PRESS) to make comparisons. For MgO, SiO2, Fe2O3, CaO, and MnO, the sparse models outperform all the others except for linear SVR, while for Na2O, K2O, TiO2, and P2O5, the sparse methods produce inferior results, likely because their emission lines in this energy range have lower transition probabilities. The strong performance of the sparse methods in this study suggests that use of dimensionality-reduction techniques as a preprocessing step may improve the performance of the linear models. Nonlinear methods tend to overfit the data and predict less accurately, while the linear methods proved to be more generalizable with better predictive performance. These results are attributed to the high dimensionality of the data (6144 channels) relative to the small number of samples studied. The best-performing models were SVR-Lin for SiO2, MgO, Fe2O3, and Na2O, lasso for Al2O3, elastic net for MnO, and PLS-1 for CaO, TiO2, and K2O. Although these differences in model performance between methods were identified, most of the models produce comparable results when p ≤ 0.05 and all techniques except kNN produced statistically-indistinguishable results. It is likely that a combination of models could be used together to yield a lower total error of prediction, depending on the requirements of the user.

Kinetic microplate bioassays for relative potency of antibiotics improved by partial Least Square (PLS) regression.

PubMed

Francisco, Fabiane Lacerda; Saviano, Alessandro Morais; Almeida, Túlia de Souza Botelho; Lourenço, Felipe Rebello

2016-05-01

Microbiological assays are widely used to estimate the relative potencies of antibiotics in order to guarantee the efficacy, safety, and quality of drug products. Despite of the advantages of turbidimetric bioassays when compared to other methods, it has limitations concerning the linearity and range of the dose-response curve determination. Here, we proposed to use partial least squares (PLS) regression to solve these limitations and to improve the prediction of relative potencies of antibiotics. Kinetic-reading microplate turbidimetric bioassays for apramacyin and vancomycin were performed using Escherichia coli (ATCC 8739) and Bacillus subtilis (ATCC 6633), respectively. Microbial growths were measured as absorbance up to 180 and 300min for apramycin and vancomycin turbidimetric bioassays, respectively. Conventional dose-response curves (absorbances or area under the microbial growth curve vs. log of antibiotic concentration) showed significant regression, however there were significant deviation of linearity. Thus, they could not be used for relative potency estimations. PLS regression allowed us to construct a predictive model for estimating the relative potencies of apramycin and vancomycin without over-fitting and it improved the linear range of turbidimetric bioassay. In addition, PLS regression provided predictions of relative potencies equivalent to those obtained from agar diffusion official methods. Therefore, we conclude that PLS regression may be used to estimate the relative potencies of antibiotics with significant advantages when compared to conventional dose-response curve determination. Copyright © 2016 Elsevier B.V. All rights reserved.

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).

Distribution of Hydroxyl Groups in Kukersite Shale Oil: Quantitative Determination Using Fourier Transform Infrared (FT-IR) Spectroscopy.

PubMed

Baird, Zachariah Steven; Oja, Vahur; Järvik, Oliver

2015-05-01

This article describes the use of Fourier transform infrared (FT-IR) spectroscopy to quantitatively measure the hydroxyl concentrations among narrow boiling shale oil cuts. Shale oil samples were from an industrial solid heat carrier retort. Reference values were measured by titration and were used to create a partial least squares regression model from FT-IR data. The model had a root mean squared error (RMSE) of 0.44 wt% OH. This method was then used to study the distribution of hydroxyl groups among more than 100 shale oil cuts, which showed that hydroxyl content increased with the average boiling point of the cut up to about 350 °C and then leveled off and decreased.

Daily Suspended Sediment Discharge Prediction Using Multiple Linear Regression and Artificial Neural Network

NASA Astrophysics Data System (ADS)

Uca; Toriman, Ekhwan; Jaafar, Othman; Maru, Rosmini; Arfan, Amal; Saleh Ahmar, Ansari

2018-01-01

Prediction of suspended sediment discharge in a catchments area is very important because it can be used to evaluation the erosion hazard, management of its water resources, water quality, hydrology project management (dams, reservoirs, and irrigation) and to determine the extent of the damage that occurred in the catchments. Multiple Linear Regression analysis and artificial neural network can be used to predict the amount of daily suspended sediment discharge. Regression analysis using the least square method, whereas artificial neural networks using Radial Basis Function (RBF) and feedforward multilayer perceptron with three learning algorithms namely Levenberg-Marquardt (LM), Scaled Conjugate Descent (SCD) and Broyden-Fletcher-Goldfarb-Shanno Quasi-Newton (BFGS). The number neuron of hidden layer is three to sixteen, while in output layer only one neuron because only one output target. The mean absolute error (MAE), root mean square error (RMSE), coefficient of determination (R2 ) and coefficient of efficiency (CE) of the multiple linear regression (MLRg) value Model 2 (6 input variable independent) has the lowest the value of MAE and RMSE (0.0000002 and 13.6039) and highest R2 and CE (0.9971 and 0.9971). When compared between LM, SCG and RBF, the BFGS model structure 3-7-1 is the better and more accurate to prediction suspended sediment discharge in Jenderam catchment. The performance value in testing process, MAE and RMSE (13.5769 and 17.9011) is smallest, meanwhile R2 and CE (0.9999 and 0.9998) is the highest if it compared with the another BFGS Quasi-Newton model (6-3-1, 9-10-1 and 12-12-1). Based on the performance statistics value, MLRg, LM, SCG, BFGS and RBF suitable and accurately for prediction by modeling the non-linear complex behavior of suspended sediment responses to rainfall, water depth and discharge. The comparison between artificial neural network (ANN) and MLRg, the MLRg Model 2 accurately for to prediction suspended sediment discharge (kg/day) in Jenderan catchment area.

[Influence of Spectral Pre-Processing on PLS Quantitative Model of Detecting Cu in Navel Orange by LIBS].

PubMed

Li, Wen-bing; Yao, Lin-tao; Liu, Mu-hua; Huang, Lin; Yao, Ming-yin; Chen, Tian-bing; He, Xiu-wen; Yang, Ping; Hu, Hui-qin; Nie, Jiang-hui

2015-05-01

Cu in navel orange was detected rapidly by laser-induced breakdown spectroscopy (LIBS) combined with partial least squares (PLS) for quantitative analysis, then the effect on the detection accuracy of the model with different spectral data ptetreatment methods was explored. Spectral data for the 52 Gannan navel orange samples were pretreated by different data smoothing, mean centralized and standard normal variable transform. Then 319~338 nm wavelength section containing characteristic spectral lines of Cu was selected to build PLS models, the main evaluation indexes of models such as regression coefficient (r), root mean square error of cross validation (RMSECV) and the root mean square error of prediction (RMSEP) were compared and analyzed. Three indicators of PLS model after 13 points smoothing and processing of the mean center were found reaching 0. 992 8, 3. 43 and 3. 4 respectively, the average relative error of prediction model is only 5. 55%, and in one word, the quality of calibration and prediction of this model are the best results. The results show that selecting the appropriate data pre-processing method, the prediction accuracy of PLS quantitative model of fruits and vegetables detected by LIBS can be improved effectively, providing a new method for fast and accurate detection of fruits and vegetables by LIBS.

Successive Projections Algorithm-Multivariable Linear Regression Classifier for the Detection of Contaminants on Chicken Carcasses in Hyperspectral Images

NASA Astrophysics Data System (ADS)

Wu, W.; Chen, G. Y.; Kang, R.; Xia, J. C.; Huang, Y. P.; Chen, K. J.

2017-07-01

During slaughtering and further processing, chicken carcasses are inevitably contaminated by microbial pathogen contaminants. Due to food safety concerns, many countries implement a zero-tolerance policy that forbids the placement of visibly contaminated carcasses in ice-water chiller tanks during processing. Manual detection of contaminants is labor consuming and imprecise. Here, a successive projections algorithm (SPA)-multivariable linear regression (MLR) classifier based on an optimal performance threshold was developed for automatic detection of contaminants on chicken carcasses. Hyperspectral images were obtained using a hyperspectral imaging system. A regression model of the classifier was established by MLR based on twelve characteristic wavelengths (505, 537, 561, 562, 564, 575, 604, 627, 656, 665, 670, and 689 nm) selected by SPA , and the optimal threshold T = 1 was obtained from the receiver operating characteristic (ROC) analysis. The SPA-MLR classifier provided the best detection results when compared with the SPA-partial least squares (PLS) regression classifier and the SPA-least squares supported vector machine (LS-SVM) classifier. The true positive rate (TPR) of 100% and the false positive rate (FPR) of 0.392% indicate that the SPA-MLR classifier can utilize spatial and spectral information to effectively detect contaminants on chicken carcasses.

Noncontact analysis of the fiber weight per unit area in prepreg by near-infrared spectroscopy.

PubMed

Jiang, B; Huang, Y D

2008-05-26

The fiber weight per unit area in prepreg is an important factor to ensure the quality of the composite products. Near-infrared spectroscopy (NIRS) technology together with a noncontact reflectance sources has been applied for quality analysis of the fiber weight per unit area. The range of the unit area fiber weight was 13.39-14.14mgcm(-2). The regression method was employed by partial least squares (PLS) and principal components regression (PCR). The calibration model was developed by 55 samples to determine the fiber weight per unit area in prepreg. The determination coefficient (R(2)), root mean square error of calibration (RMSEC) and root mean square error of prediction (RMSEP) were 0.82, 0.092, 0.099, respectively. The predicted values of the fiber weight per unit area in prepreg measured by NIRS technology were comparable to the values obtained by the reference method. For this technology, the noncontact reflectance sources focused directly on the sample with neither previous treatment nor manipulation. The results of the paired t-test revealed that there was no significant difference between the NIR method and the reference method. Besides, the prepreg could be analyzed one time within 20s without sample destruction.

Statistical downscaling of precipitation using long short-term memory recurrent neural networks

NASA Astrophysics Data System (ADS)

Misra, Saptarshi; Sarkar, Sudeshna; Mitra, Pabitra

2017-11-01

Hydrological impacts of global climate change on regional scale are generally assessed by downscaling large-scale climatic variables, simulated by General Circulation Models (GCMs), to regional, small-scale hydrometeorological variables like precipitation, temperature, etc. In this study, we propose a new statistical downscaling model based on Recurrent Neural Network with Long Short-Term Memory which captures the spatio-temporal dependencies in local rainfall. The previous studies have used several other methods such as linear regression, quantile regression, kernel regression, beta regression, and artificial neural networks. Deep neural networks and recurrent neural networks have been shown to be highly promising in modeling complex and highly non-linear relationships between input and output variables in different domains and hence we investigated their performance in the task of statistical downscaling. We have tested this model on two datasets—one on precipitation in Mahanadi basin in India and the second on precipitation in Campbell River basin in Canada. Our autoencoder coupled long short-term memory recurrent neural network model performs the best compared to other existing methods on both the datasets with respect to temporal cross-correlation, mean squared error, and capturing the extremes.

Effects of land cover, topography, and built structure on seasonal water quality at multiple spatial scales.

PubMed

Pratt, Bethany; Chang, Heejun

2012-03-30

The relationship among land cover, topography, built structure and stream water quality in the Portland Metro region of Oregon and Clark County, Washington areas, USA, is analyzed using ordinary least squares (OLS) and geographically weighted (GWR) multiple regression models. Two scales of analysis, a sectional watershed and a buffer, offered a local and a global investigation of the sources of stream pollutants. Model accuracy, measured by R(2) values, fluctuated according to the scale, season, and regression method used. While most wet season water quality parameters are associated with urban land covers, most dry season water quality parameters are related topographic features such as elevation and slope. GWR models, which take into consideration local relations of spatial autocorrelation, had stronger results than OLS regression models. In the multiple regression models, sectioned watershed results were consistently better than the sectioned buffer results, except for dry season pH and stream temperature parameters. This suggests that while riparian land cover does have an effect on water quality, a wider contributing area needs to be included in order to account for distant sources of pollutants. Copyright © 2012 Elsevier B.V. All rights reserved.

Comparison of methods for the analysis of relatively simple mediation models.

PubMed

Rijnhart, Judith J M; Twisk, Jos W R; Chinapaw, Mai J M; de Boer, Michiel R; Heymans, Martijn W

2017-09-01

Statistical mediation analysis is an often used method in trials, to unravel the pathways underlying the effect of an intervention on a particular outcome variable. Throughout the years, several methods have been proposed, such as ordinary least square (OLS) regression, structural equation modeling (SEM), and the potential outcomes framework. Most applied researchers do not know that these methods are mathematically equivalent when applied to mediation models with a continuous mediator and outcome variable. Therefore, the aim of this paper was to demonstrate the similarities between OLS regression, SEM, and the potential outcomes framework in three mediation models: 1) a crude model, 2) a confounder-adjusted model, and 3) a model with an interaction term for exposure-mediator interaction. Secondary data analysis of a randomized controlled trial that included 546 schoolchildren. In our data example, the mediator and outcome variable were both continuous. We compared the estimates of the total, direct and indirect effects, proportion mediated, and 95% confidence intervals (CIs) for the indirect effect across OLS regression, SEM, and the potential outcomes framework. OLS regression, SEM, and the potential outcomes framework yielded the same effect estimates in the crude mediation model, the confounder-adjusted mediation model, and the mediation model with an interaction term for exposure-mediator interaction. Since OLS regression, SEM, and the potential outcomes framework yield the same results in three mediation models with a continuous mediator and outcome variable, researchers can continue using the method that is most convenient to them.

Evaluation of the CEAS model for barley yields in North Dakota and Minnesota

NASA Technical Reports Server (NTRS)

Barnett, T. L. (Principal Investigator)

1981-01-01

The CEAS yield model is based upon multiple regression analysis at the CRD and state levels. For the historical time series, yield is regressed on a set of variables derived from monthly mean temperature and monthly precipitation. Technological trend is represented by piecewise linear and/or quadriatic functions of year. Indicators of yield reliability obtained from a ten-year bootstrap test (1970-79) demonstrated that biases are small and performance as indicated by the root mean square errors are acceptable for intended application, however, model response for individual years particularly unusual years, is not very reliable and shows some large errors. The model is objective, adequate, timely, simple and not costly. It considers scientific knowledge on a broad scale but not in detail, and does not provide a good current measure of modeled yield reliability.

An Interactive Computer Package for Use with Simulation Models Which Performs Multidimensional Sensitivity Analysis by Employing the Techniques of Response Surface Methodology.

DTIC Science & Technology

1984-12-01

total sum of squares at the center points minus the correction factor for the mean at the center points ( SSpe =Y’Y-nlY), where n1 is the number of...SSlac=SSres- SSpe ). The sum of squares due to pure error estimates 0" and the sum of squares due to lack-of-fit estimates 0’" plus a bias term if...Response Surface Methodology Source d.f. SS MS Regression n b’X1 Y b’XVY/n Residual rn-n Y’Y-b’X’ *Y (Y’Y-b’X’Y)/(n-n) Pure Error ni-i Y’Y-nl1Y SSpe / (ni

Analysis of spreadable cheese by Raman spectroscopy and chemometric tools.

PubMed

Oliveira, Kamila de Sá; Callegaro, Layce de Souza; Stephani, Rodrigo; Almeida, Mariana Ramos; de Oliveira, Luiz Fernando Cappa

2016-03-01

In this work, FT-Raman spectroscopy was explored to evaluate spreadable cheese samples. A partial least squares discriminant analysis was employed to identify the spreadable cheese samples containing starch. To build the models, two types of samples were used: commercial samples and samples manufactured in local industries. The method of supervised classification PLS-DA was employed to classify the samples as adulterated or without starch. Multivariate regression was performed using the partial least squares method to quantify the starch in the spreadable cheese. The limit of detection obtained for the model was 0.34% (w/w) and the limit of quantification was 1.14% (w/w). The reliability of the models was evaluated by determining the confidence interval, which was calculated using the bootstrap re-sampling technique. The results show that the classification models can be used to complement classical analysis and as screening methods. Copyright © 2015 Elsevier Ltd. All rights reserved.

Diagnosing and dealing with multicollinearity.

PubMed

Schroeder, M A

1990-04-01

The purpose of this article was to increase nurse researchers' awareness of the effects of collinear data in developing theoretical models for nursing practice. Collinear data distort the true value of the estimates generated from ordinary least-squares analysis. Theoretical models developed to provide the underpinnings of nursing practice need not be abandoned, however, because they fail to produce consistent estimates over repeated applications. It is also important to realize that multicollinearity is a data problem, not a problem associated with misspecification of a theorectical model. An investigator must first be aware of the problem, and then it is possible to develop an educated solution based on the degree of multicollinearity, theoretical considerations, and sources of error associated with alternative, biased, least-square regression techniques. Decisions based on theoretical and statistical considerations will further the development of theory-based nursing practice.

Tire-road friction coefficient estimation based on the resonance frequency of in-wheel motor drive system

NASA Astrophysics Data System (ADS)

Chen, Long; Bian, Mingyuan; Luo, Yugong; Qin, Zhaobo; Li, Keqiang

2016-01-01

In this paper, a resonance frequency-based tire-road friction coefficient (TRFC) estimation method is proposed by considering the dynamics performance of the in-wheel motor drive system under small slip ratio conditions. A frequency response function (FRF) is deduced for the drive system that is composed of a dynamic tire model and a simplified motor model. A linear relationship between the squared system resonance frequency and the TFRC is described with the FRF. Furthermore, the resonance frequency is identified by the Auto-Regressive eXogenous model using the information of the motor torque and the wheel speed, and the TRFC is estimated thereafter by a recursive least squares filter with the identified resonance frequency. Finally, the effectiveness of the proposed approach is demonstrated through simulations and experimental tests on different road surfaces.

Hyperspectral Imaging in Tandem with R Statistics and Image Processing for Detection and Visualization of pH in Japanese Big Sausages Under Different Storage Conditions.

PubMed

Feng, Chao-Hui; Makino, Yoshio; Yoshimura, Masatoshi; Thuyet, Dang Quoc; García-Martín, Juan Francisco

2018-02-01

The potential of hyperspectral imaging with wavelengths of 380 to 1000 nm was used to determine the pH of cooked sausages after different storage conditions (4 °C for 1 d, 35 °C for 1, 3, and 5 d). The mean spectra of the sausages were extracted from the hyperspectral images and partial least squares regression (PLSR) model was developed to relate spectral profiles with the pH of the cooked sausages. Eleven important wavelengths were selected based on the regression coefficient values. The PLSR model established using the optimal wavelengths showed good precision being the prediction coefficient of determination (R p 2 ) 0.909 and the root mean square error of prediction 0.035. The prediction map for illustrating pH indices in sausages was for the first time developed by R statistics. The overall results suggested that hyperspectral imaging combined with PLSR and R statistics are capable to quantify and visualize the sausages pH evolution under different storage conditions. In this paper, hyperspectral imaging is for the first time used to detect pH in cooked sausages using R statistics, which provides another useful information for the researchers who do not have the access to Matlab. Eleven optimal wavelengths were successfully selected, which were used for simplifying the PLSR model established based on the full wavelengths. This simplified model achieved a high R p 2 (0.909) and a low root mean square error of prediction (0.035), which can be useful for the design of multispectral imaging systems. © 2017 Institute of Food Technologists®.

Rapid Quantitative Determination of Squalene in Shark Liver Oils by Raman and IR Spectroscopy.

PubMed

Hall, David W; Marshall, Susan N; Gordon, Keith C; Killeen, Daniel P

2016-01-01

Squalene is sourced predominantly from shark liver oils and to a lesser extent from plants such as olives. It is used for the production of surfactants, dyes, sunscreen, and cosmetics. The economic value of shark liver oil is directly related to the squalene content, which in turn is highly variable and species-dependent. Presented here is a validated gas chromatography-mass spectrometry analysis method for the quantitation of squalene in shark liver oils, with an accuracy of 99.0 %, precision of 0.23 % (standard deviation), and linearity of >0.999. The method has been used to measure the squalene concentration of 16 commercial shark liver oils. These reference squalene concentrations were related to infrared (IR) and Raman spectra of the same oils using partial least squares regression. The resultant models were suitable for the rapid quantitation of squalene in shark liver oils, with cross-validation r (2) values of >0.98 and root mean square errors of validation of ≤4.3 % w/w. Independent test set validation of these models found mean absolute deviations of the 4.9 and 1.0 % w/w for the IR and Raman models, respectively. Both techniques were more accurate than results obtained by an industrial refractive index analysis method, which is used for rapid, cheap quantitation of squalene in shark liver oils. In particular, the Raman partial least squares regression was suited to quantitative squalene analysis. The intense and highly characteristic Raman bands of squalene made quantitative analysis possible irrespective of the lipid matrix.

Rapid determination of crocins in saffron by near-infrared spectroscopy combined with chemometric techniques

NASA Astrophysics Data System (ADS)

Li, Shuailing; Shao, Qingsong; Lu, Zhonghua; Duan, Chengli; Yi, Haojun; Su, Liyang

2018-02-01

Saffron is an expensive spice. Its primary effective constituents are crocin I and II, and the contents of these compounds directly affect the quality and commercial value of saffron. In this study, near-infrared spectroscopy was combined with chemometric techniques for the determination of crocin I and II in saffron. Partial least squares regression models were built for the quantification of crocin I and II. By comparing different spectral ranges and spectral pretreatment methods (no pretreatment, vector normalization, subtract a straight line, multiplicative scatter correction, minimum-maximum normalization, eliminate the constant offset, first derivative, and second derivative), optimum models were developed. The root mean square error of cross-validation values of the best partial least squares models for crocin I and II were 1.40 and 0.30, respectively. The coefficients of determination for crocin I and II were 93.40 and 96.30, respectively. These results show that near-infrared spectroscopy can be combined with chemometric techniques to determine the contents of crocin I and II in saffron quickly and efficiently.

Fast detection and visualization of minced lamb meat adulteration using NIR hyperspectral imaging and multivariate image analysis.

PubMed

Kamruzzaman, Mohammed; Sun, Da-Wen; ElMasry, Gamal; Allen, Paul

2013-01-15

Many studies have been carried out in developing non-destructive technologies for predicting meat adulteration, but there is still no endeavor for non-destructive detection and quantification of adulteration in minced lamb meat. The main goal of this study was to develop and optimize a rapid analytical technique based on near-infrared (NIR) hyperspectral imaging to detect the level of adulteration in minced lamb. Initial investigation was carried out using principal component analysis (PCA) to identify the most potential adulterate in minced lamb. Minced lamb meat samples were then adulterated with minced pork in the range 2-40% (w/w) at approximately 2% increments. Spectral data were used to develop a partial least squares regression (PLSR) model to predict the level of adulteration in minced lamb. Good prediction model was obtained using the whole spectral range (910-1700 nm) with a coefficient of determination (R(2)(cv)) of 0.99 and root-mean-square errors estimated by cross validation (RMSECV) of 1.37%. Four important wavelengths (940, 1067, 1144 and 1217 nm) were selected using weighted regression coefficients (Bw) and a multiple linear regression (MLR) model was then established using these important wavelengths to predict adulteration. The MLR model resulted in a coefficient of determination (R(2)(cv)) of 0.98 and RMSECV of 1.45%. The developed MLR model was then applied to each pixel in the image to obtain prediction maps to visualize the distribution of adulteration of the tested samples. The results demonstrated that the laborious and time-consuming tradition analytical techniques could be replaced by spectral data in order to provide rapid, low cost and non-destructive testing technique for adulterate detection in minced lamb meat. Copyright © 2012 Elsevier B.V. All rights reserved.

Measurement of process variables in solid-state fermentation of wheat straw using FT-NIR spectroscopy and synergy interval PLS algorithm

NASA Astrophysics Data System (ADS)

Jiang, Hui; Liu, Guohai; Mei, Congli; Yu, Shuang; Xiao, Xiahong; Ding, Yuhan

2012-11-01

The feasibility of rapid determination of the process variables (i.e. pH and moisture content) in solid-state fermentation (SSF) of wheat straw using Fourier transform near infrared (FT-NIR) spectroscopy was studied. Synergy interval partial least squares (siPLS) algorithm was implemented to calibrate regression model. The number of PLS factors and the number of subintervals were optimized simultaneously by cross-validation. The performance of the prediction model was evaluated according to the root mean square error of cross-validation (RMSECV), the root mean square error of prediction (RMSEP) and the correlation coefficient (R). The measurement results of the optimal model were obtained as follows: RMSECV = 0.0776, Rc = 0.9777, RMSEP = 0.0963, and Rp = 0.9686 for pH model; RMSECV = 1.3544% w/w, Rc = 0.8871, RMSEP = 1.4946% w/w, and Rp = 0.8684 for moisture content model. Finally, compared with classic PLS and iPLS models, the siPLS model revealed its superior performance. The overall results demonstrate that FT-NIR spectroscopy combined with siPLS algorithm can be used to measure process variables in solid-state fermentation of wheat straw, and NIR spectroscopy technique has a potential to be utilized in SSF industry.

Use of Thematic Mapper for water quality assessment

NASA Technical Reports Server (NTRS)

Horn, E. M.; Morrissey, L. A.

1984-01-01

The evaluation of simulated TM data obtained on an ER-2 aircraft at twenty-five predesignated sample sites for mapping water quality factors such as conductivity, pH, suspended solids, turbidity, temperature, and depth, is discussed. Using a multiple regression for the seven TM bands, an equation is developed for the suspended solids. TM bands 1, 2, 3, 4, and 6 are used with logarithm conductivity in a multiple regression. The assessment of regression equations for a high coefficient of determination (R-squared) and statistical significance is considered. Confidence intervals about the mean regression point are calculated in order to assess the robustness of the regressions used for mapping conductivity, turbidity, and suspended solids, and by regressing random subsamples of sites and comparing the resultant range of R-squared, cross validation is conducted.

Raman spectroscopy-based screening of hepatitis C and associated molecular changes

NASA Astrophysics Data System (ADS)

Bilal, Maria; Bilal, M.; Saleem, M.; Khan, Saranjam; Ullah, Rahat; Fatima, Kiran; Ahmed, M.; Hayat, Abbas; Shahzada, Shaista; Ullah Khan, Ehsan

2017-09-01

This study presents the optical screening of hepatitis C and its associated molecular changes in human blood sera using a partial least-squares regression model based on their Raman spectra. In total, 152 samples were tested through enzyme-linked immunosorbent assay for confirmation. This model utilizes minor spectral variations in the Raman spectra of the positive and control groups. Regression coefficients of this model were analyzed with reference to the variations in concentration of associated molecules in these two groups. It was found that trehalose, chitin, ammonia, and cytokines are positively correlated while lipids, beta structures of proteins, and carbohydrate-binding proteins are negatively correlated with hepatitis C. The regression vector yielded by this model is utilized to predict hepatitis C in unknown samples. This model has been evaluated by a cross-validation method, which yielded a correlation coefficient of 0.91. Moreover, 30 unknown samples were screened for hepatitis C infection using this model to test its performance. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve from these predictions were found to be 93.3%, 100%, 96.7%, and 1, respectively.

Monthly streamflow forecasting with auto-regressive integrated moving average

NASA Astrophysics Data System (ADS)

Nasir, Najah; Samsudin, Ruhaidah; Shabri, Ani

2017-09-01

Forecasting of streamflow is one of the many ways that can contribute to better decision making for water resource management. The auto-regressive integrated moving average (ARIMA) model was selected in this research for monthly streamflow forecasting with enhancement made by pre-processing the data using singular spectrum analysis (SSA). This study also proposed an extension of the SSA technique to include a step where clustering was performed on the eigenvector pairs before reconstruction of the time series. The monthly streamflow data of Sungai Muda at Jeniang, Sungai Muda at Jambatan Syed Omar and Sungai Ketil at Kuala Pegang was gathered from the Department of Irrigation and Drainage Malaysia. A ratio of 9:1 was used to divide the data into training and testing sets. The ARIMA, SSA-ARIMA and Clustered SSA-ARIMA models were all developed in R software. Results from the proposed model are then compared to a conventional auto-regressive integrated moving average model using the root-mean-square error and mean absolute error values. It was found that the proposed model can outperform the conventional model.

Modelling the breeding of Aedes Albopictus species in an urban area in Pulau Pinang using polynomial regression

NASA Astrophysics Data System (ADS)

Salleh, Nur Hanim Mohd; Ali, Zalila; Noor, Norlida Mohd.; Baharum, Adam; Saad, Ahmad Ramli; Sulaiman, Husna Mahirah; Ahmad, Wan Muhamad Amir W.

2014-07-01

Polynomial regression is used to model a curvilinear relationship between a response variable and one or more predictor variables. It is a form of a least squares linear regression model that predicts a single response variable by decomposing the predictor variables into an nth order polynomial. In a curvilinear relationship, each curve has a number of extreme points equal to the highest order term in the polynomial. A quadratic model will have either a single maximum or minimum, whereas a cubic model has both a relative maximum and a minimum. This study used quadratic modeling techniques to analyze the effects of environmental factors: temperature, relative humidity, and rainfall distribution on the breeding of Aedes albopictus, a type of Aedes mosquito. Data were collected at an urban area in south-west Penang from September 2010 until January 2011. The results indicated that the breeding of Aedes albopictus in the urban area is influenced by all three environmental characteristics. The number of mosquito eggs is estimated to reach a maximum value at a medium temperature, a medium relative humidity and a high rainfall distribution.

Textbook Readability and Student Performance in Online Introductory Corporate Finance Classes

ERIC Educational Resources Information Center

Peng, Chien-Chih

2015-01-01

This paper examines whether the choice of a more readable textbook can improve student performance in online introductory corporate finance classes. The ordinary least squares regression model is employed to analyze a sample of 206 students during the period from 2008 to 2012. The results of this study show that the student's age, student's major,…

Characteristics and Psychosocial Predictors of Adolescent Nonsuicidal Self-Injury in Residential Care

ERIC Educational Resources Information Center

Gallant, Jason; Snyder, Gregory S.; von der Embse, Nathaniel P.

2014-01-01

This study examined characteristics and biopsychosocial predictors of nonsuicidal self-injury in a sample (N = 753) of youth in residential care admitted between 2005 and 2010. To model the data, the authors used t-tests, chi-square tests, and multiple logistic regressions stratified by gender. Results suggested that 12% of youth engaged in…

Generalized Partial Least Squares Approach for Nominal Multinomial Logit Regression Models with a Functional Covariate

ERIC Educational Resources Information Center

Albaqshi, Amani Mohammed H.

2017-01-01

Functional Data Analysis (FDA) has attracted substantial attention for the last two decades. Within FDA, classifying curves into two or more categories is consistently of interest to scientists, but multi-class prediction within FDA is challenged in that most classification tools have been limited to binary response applications. The functional…

Baseline Correction of Diffuse Reflection Near-Infrared Spectra Using Searching Region Standard Normal Variate (SRSNV).

PubMed

Genkawa, Takuma; Shinzawa, Hideyuki; Kato, Hideaki; Ishikawa, Daitaro; Murayama, Kodai; Komiyama, Makoto; Ozaki, Yukihiro

2015-12-01

An alternative baseline correction method for diffuse reflection near-infrared (NIR) spectra, searching region standard normal variate (SRSNV), was proposed. Standard normal variate (SNV) is an effective pretreatment method for baseline correction of diffuse reflection NIR spectra of powder and granular samples; however, its baseline correction performance depends on the NIR region used for SNV calculation. To search for an optimal NIR region for baseline correction using SNV, SRSNV employs moving window partial least squares regression (MWPLSR), and an optimal NIR region is identified based on the root mean square error (RMSE) of cross-validation of the partial least squares regression (PLSR) models with the first latent variable (LV). The performance of SRSNV was evaluated using diffuse reflection NIR spectra of mixture samples consisting of wheat flour and granular glucose (0-100% glucose at 5% intervals). From the obtained NIR spectra of the mixture in the 10 000-4000 cm(-1) region at 4 cm intervals (1501 spectral channels), a series of spectral windows consisting of 80 spectral channels was constructed, and then SNV spectra were calculated for each spectral window. Using these SNV spectra, a series of PLSR models with the first LV for glucose concentration was built. A plot of RMSE versus the spectral window position obtained using the PLSR models revealed that the 8680–8364 cm(-1) region was optimal for baseline correction using SNV. In the SNV spectra calculated using the 8680–8364 cm(-1) region (SRSNV spectra), a remarkable relative intensity change between a band due to wheat flour at 8500 cm(-1) and that due to glucose at 8364 cm(-1) was observed owing to successful baseline correction using SNV. A PLSR model with the first LV based on the SRSNV spectra yielded a determination coefficient (R2) of 0.999 and an RMSE of 0.70%, while a PLSR model with three LVs based on SNV spectra calculated in the full spectral region gave an R2 of 0.995 and an RMSE of 2.29%. Additional evaluation of SRSNV was carried out using diffuse reflection NIR spectra of marzipan and corn samples, and PLSR models based on SRSNV spectra showed good prediction results. These evaluation results indicate that SRSNV is effective in baseline correction of diffuse reflection NIR spectra and provides regression models with good prediction accuracy.

A land use regression model for ambient ultrafine particles in Montreal, Canada: A comparison of linear regression and a machine learning approach.

PubMed

Weichenthal, Scott; Ryswyk, Keith Van; Goldstein, Alon; Bagg, Scott; Shekkarizfard, Maryam; Hatzopoulou, Marianne

2016-04-01

Existing evidence suggests that ambient ultrafine particles (UFPs) (<0.1µm) may contribute to acute cardiorespiratory morbidity. However, few studies have examined the long-term health effects of these pollutants owing in part to a need for exposure surfaces that can be applied in large population-based studies. To address this need, we developed a land use regression model for UFPs in Montreal, Canada using mobile monitoring data collected from 414 road segments during the summer and winter months between 2011 and 2012. Two different approaches were examined for model development including standard multivariable linear regression and a machine learning approach (kernel-based regularized least squares (KRLS)) that learns the functional form of covariate impacts on ambient UFP concentrations from the data. The final models included parameters for population density, ambient temperature and wind speed, land use parameters (park space and open space), length of local roads and rail, and estimated annual average NOx emissions from traffic. The final multivariable linear regression model explained 62% of the spatial variation in ambient UFP concentrations whereas the KRLS model explained 79% of the variance. The KRLS model performed slightly better than the linear regression model when evaluated using an external dataset (R(2)=0.58 vs. 0.55) or a cross-validation procedure (R(2)=0.67 vs. 0.60). In general, our findings suggest that the KRLS approach may offer modest improvements in predictive performance compared to standard multivariable linear regression models used to estimate spatial variations in ambient UFPs. However, differences in predictive performance were not statistically significant when evaluated using the cross-validation procedure. Crown Copyright © 2015. Published by Elsevier Inc. All rights reserved.

Application of Rapid Visco Analyser (RVA) viscograms and chemometrics for maize hardness characterisation.

PubMed

Guelpa, Anina; Bevilacqua, Marta; Marini, Federico; O'Kennedy, Kim; Geladi, Paul; Manley, Marena

2015-04-15

It has been established in this study that the Rapid Visco Analyser (RVA) can describe maize hardness, irrespective of the RVA profile, when used in association with appropriate multivariate data analysis techniques. Therefore, the RVA can complement or replace current and/or conventional methods as a hardness descriptor. Hardness modelling based on RVA viscograms was carried out using seven conventional hardness methods (hectoliter mass (HLM), hundred kernel mass (HKM), particle size index (PSI), percentage vitreous endosperm (%VE), protein content, percentage chop (%chop) and near infrared (NIR) spectroscopy) as references and three different RVA profiles (hard, soft and standard) as predictors. An approach using locally weighted partial least squares (LW-PLS) was followed to build the regression models. The resulted prediction errors (root mean square error of cross-validation (RMSECV) and root mean square error of prediction (RMSEP)) for the quantification of hardness values were always lower or in the same order of the laboratory error of the reference method. Copyright © 2014 Elsevier Ltd. All rights reserved.

Application of Fourier transform infrared spectroscopy and orthogonal projections to latent structures/partial least squares regression for estimation of procyanidins average degree of polymerisation.

PubMed

Passos, Cláudia P; Cardoso, Susana M; Barros, António S; Silva, Carlos M; Coimbra, Manuel A

2010-02-28

Fourier transform infrared (FTIR) spectroscopy has being emphasised as a widespread technique in the quick assess of food components. In this work, procyanidins were extracted with methanol and acetone/water from the seeds of white and red grape varieties. A fractionation by graded methanol/chloroform precipitations allowed to obtain 26 samples that were characterised using thiolysis as pre-treatment followed by HPLC-UV and MS detection. The average degree of polymerisation (DPn) of the procyanidins in the samples ranged from 2 to 11 flavan-3-ol residues. FTIR spectroscopy within the wavenumbers region of 1800-700 cm(-1) allowed to build a partial least squares (PLS1) regression model with 8 latent variables (LVs) for the estimation of the DPn, giving a RMSECV of 11.7%, with a R(2) of 0.91 and a RMSEP of 2.58. The application of orthogonal projection to latent structures (O-PLS1) clarifies the interpretation of the regression model vectors. Moreover, the O-PLS procedure has removed 88% of non-correlated variations with the DPn, allowing to relate the increase of the absorbance peaks at 1203 and 1099 cm(-1) with the increase of the DPn due to the higher proportion of substitutions in the aromatic ring of the polymerised procyanidin molecules. Copyright 2009 Elsevier B.V. All rights reserved.

Revisiting Gaussian Process Regression Modeling for Localization in Wireless Sensor Networks

PubMed Central

Richter, Philipp; Toledano-Ayala, Manuel

2015-01-01

Signal strength-based positioning in wireless sensor networks is a key technology for seamless, ubiquitous localization, especially in areas where Global Navigation Satellite System (GNSS) signals propagate poorly. To enable wireless local area network (WLAN) location fingerprinting in larger areas while maintaining accuracy, methods to reduce the effort of radio map creation must be consolidated and automatized. Gaussian process regression has been applied to overcome this issue, also with auspicious results, but the fit of the model was never thoroughly assessed. Instead, most studies trained a readily available model, relying on the zero mean and squared exponential covariance function, without further scrutinization. This paper studies the Gaussian process regression model selection for WLAN fingerprinting in indoor and outdoor environments. We train several models for indoor/outdoor- and combined areas; we evaluate them quantitatively and compare them by means of adequate model measures, hence assessing the fit of these models directly. To illuminate the quality of the model fit, the residuals of the proposed model are investigated, as well. Comparative experiments on the positioning performance verify and conclude the model selection. In this way, we show that the standard model is not the most appropriate, discuss alternatives and present our best candidate. PMID:26370996

Time series forecasting using ERNN and QR based on Bayesian model averaging

NASA Astrophysics Data System (ADS)

Pwasong, Augustine; Sathasivam, Saratha

2017-08-01

The Bayesian model averaging technique is a multi-model combination technique. The technique was employed to amalgamate the Elman recurrent neural network (ERNN) technique with the quadratic regression (QR) technique. The amalgamation produced a hybrid technique known as the hybrid ERNN-QR technique. The potentials of forecasting with the hybrid technique are compared with the forecasting capabilities of individual techniques of ERNN and QR. The outcome revealed that the hybrid technique is superior to the individual techniques in the mean square error sense.

Modeling the human development index and the percentage of poor people using quantile smoothing splines

NASA Astrophysics Data System (ADS)

Mulyani, Sri; Andriyana, Yudhie; Sudartianto

2017-03-01

Mean regression is a statistical method to explain the relationship between the response variable and the predictor variable based on the central tendency of the data (mean) of the response variable. The parameter estimation in mean regression (with Ordinary Least Square or OLS) generates a problem if we apply it to the data with a symmetric, fat-tailed, or containing outlier. Hence, an alternative method is necessary to be used to that kind of data, for example quantile regression method. The quantile regression is a robust technique to the outlier. This model can explain the relationship between the response variable and the predictor variable, not only on the central tendency of the data (median) but also on various quantile, in order to obtain complete information about that relationship. In this study, a quantile regression is developed with a nonparametric approach such as smoothing spline. Nonparametric approach is used if the prespecification model is difficult to determine, the relation between two variables follow the unknown function. We will apply that proposed method to poverty data. Here, we want to estimate the Percentage of Poor People as the response variable involving the Human Development Index (HDI) as the predictor variable.

Cervical Vertebral Body's Volume as a New Parameter for Predicting the Skeletal Maturation Stages.

PubMed

Choi, Youn-Kyung; Kim, Jinmi; Yamaguchi, Tetsutaro; Maki, Koutaro; Ko, Ching-Chang; Kim, Yong-Il

2016-01-01

This study aimed to determine the correlation between the volumetric parameters derived from the images of the second, third, and fourth cervical vertebrae by using cone beam computed tomography with skeletal maturation stages and to propose a new formula for predicting skeletal maturation by using regression analysis. We obtained the estimation of skeletal maturation levels from hand-wrist radiographs and volume parameters derived from the second, third, and fourth cervical vertebrae bodies from 102 Japanese patients (54 women and 48 men, 5-18 years of age). We performed Pearson's correlation coefficient analysis and simple regression analysis. All volume parameters derived from the second, third, and fourth cervical vertebrae exhibited statistically significant correlations (P < 0.05). The simple regression model with the greatest R-square indicated the fourth-cervical-vertebra volume as an independent variable with a variance inflation factor less than ten. The explanation power was 81.76%. Volumetric parameters of cervical vertebrae using cone beam computed tomography are useful in regression models. The derived regression model has the potential for clinical application as it enables a simple and quantitative analysis to evaluate skeletal maturation level.

Cervical Vertebral Body's Volume as a New Parameter for Predicting the Skeletal Maturation Stages

PubMed Central

Choi, Youn-Kyung; Kim, Jinmi; Maki, Koutaro; Ko, Ching-Chang

2016-01-01

This study aimed to determine the correlation between the volumetric parameters derived from the images of the second, third, and fourth cervical vertebrae by using cone beam computed tomography with skeletal maturation stages and to propose a new formula for predicting skeletal maturation by using regression analysis. We obtained the estimation of skeletal maturation levels from hand-wrist radiographs and volume parameters derived from the second, third, and fourth cervical vertebrae bodies from 102 Japanese patients (54 women and 48 men, 5–18 years of age). We performed Pearson's correlation coefficient analysis and simple regression analysis. All volume parameters derived from the second, third, and fourth cervical vertebrae exhibited statistically significant correlations (P < 0.05). The simple regression model with the greatest R-square indicated the fourth-cervical-vertebra volume as an independent variable with a variance inflation factor less than ten. The explanation power was 81.76%. Volumetric parameters of cervical vertebrae using cone beam computed tomography are useful in regression models. The derived regression model has the potential for clinical application as it enables a simple and quantitative analysis to evaluate skeletal maturation level. PMID:27340668

Pragmatic estimation of a spatio-temporal air quality model with irregular monitoring data

NASA Astrophysics Data System (ADS)

Sampson, Paul D.; Szpiro, Adam A.; Sheppard, Lianne; Lindström, Johan; Kaufman, Joel D.

2011-11-01

Statistical analyses of health effects of air pollution have increasingly used GIS-based covariates for prediction of ambient air quality in "land use" regression models. More recently these spatial regression models have accounted for spatial correlation structure in combining monitoring data with land use covariates. We present a flexible spatio-temporal modeling framework and pragmatic, multi-step estimation procedure that accommodates essentially arbitrary patterns of missing data with respect to an ideally complete space by time matrix of observations on a network of monitoring sites. The methodology incorporates a model for smooth temporal trends with coefficients varying in space according to Partial Least Squares regressions on a large set of geographic covariates and nonstationary modeling of spatio-temporal residuals from these regressions. This work was developed to provide spatial point predictions of PM 2.5 concentrations for the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air) using irregular monitoring data derived from the AQS regulatory monitoring network and supplemental short-time scale monitoring campaigns conducted to better predict intra-urban variation in air quality. We demonstrate the interpretation and accuracy of this methodology in modeling data from 2000 through 2006 in six U.S. metropolitan areas and establish a basis for likelihood-based estimation.

Legitimate Techniques for Improving the R-Square and Related Statistics of a Multiple Regression Model

DTIC Science & Technology

1981-01-01

explanatory variable has been ommitted. Ramsey (1974) has developed a rather interesting test for detecting specification errors using estimates of the...Peter. (1979) A Guide to Econometrics , Cambridge, MA: The MIT Press. Ramsey , J.B. (1974), "Classical Model Selection Through Specification Error... Tests ," in P. Zarembka, Ed. Frontiers in Econometrics , New York: Academia Press. Theil, Henri. (1971), Principles of Econometrics , New York: John Wiley

Kendall-Theil Robust Line (KTRLine--version 1.0)-A Visual Basic Program for Calculating and Graphing Robust Nonparametric Estimates of Linear-Regression Coefficients Between Two Continuous Variables

USGS Publications Warehouse

Granato, Gregory E.

2006-01-01

The Kendall-Theil Robust Line software (KTRLine-version 1.0) is a Visual Basic program that may be used with the Microsoft Windows operating system to calculate parameters for robust, nonparametric estimates of linear-regression coefficients between two continuous variables. The KTRLine software was developed by the U.S. Geological Survey, in cooperation with the Federal Highway Administration, for use in stochastic data modeling with local, regional, and national hydrologic data sets to develop planning-level estimates of potential effects of highway runoff on the quality of receiving waters. The Kendall-Theil robust line was selected because this robust nonparametric method is resistant to the effects of outliers and nonnormality in residuals that commonly characterize hydrologic data sets. The slope of the line is calculated as the median of all possible pairwise slopes between points. The intercept is calculated so that the line will run through the median of input data. A single-line model or a multisegment model may be specified. The program was developed to provide regression equations with an error component for stochastic data generation because nonparametric multisegment regression tools are not available with the software that is commonly used to develop regression models. The Kendall-Theil robust line is a median line and, therefore, may underestimate total mass, volume, or loads unless the error component or a bias correction factor is incorporated into the estimate. Regression statistics such as the median error, the median absolute deviation, the prediction error sum of squares, the root mean square error, the confidence interval for the slope, and the bias correction factor for median estimates are calculated by use of nonparametric methods. These statistics, however, may be used to formulate estimates of mass, volume, or total loads. The program is used to read a two- or three-column tab-delimited input file with variable names in the first row and data in subsequent rows. The user may choose the columns that contain the independent (X) and dependent (Y) variable. A third column, if present, may contain metadata such as the sample-collection location and date. The program screens the input files and plots the data. The KTRLine software is a graphical tool that facilitates development of regression models by use of graphs of the regression line with data, the regression residuals (with X or Y), and percentile plots of the cumulative frequency of the X variable, Y variable, and the regression residuals. The user may individually transform the independent and dependent variables to reduce heteroscedasticity and to linearize data. The program plots the data and the regression line. The program also prints model specifications and regression statistics to the screen. The user may save and print the regression results. The program can accept data sets that contain up to about 15,000 XY data points, but because the program must sort the array of all pairwise slopes, the program may be perceptibly slow with data sets that contain more than about 1,000 points.

Testing a single regression coefficient in high dimensional linear models

PubMed Central

Zhong, Ping-Shou; Li, Runze; Wang, Hansheng; Tsai, Chih-Ling

2017-01-01

In linear regression models with high dimensional data, the classical z-test (or t-test) for testing the significance of each single regression coefficient is no longer applicable. This is mainly because the number of covariates exceeds the sample size. In this paper, we propose a simple and novel alternative by introducing the Correlated Predictors Screening (CPS) method to control for predictors that are highly correlated with the target covariate. Accordingly, the classical ordinary least squares approach can be employed to estimate the regression coefficient associated with the target covariate. In addition, we demonstrate that the resulting estimator is consistent and asymptotically normal even if the random errors are heteroscedastic. This enables us to apply the z-test to assess the significance of each covariate. Based on the p-value obtained from testing the significance of each covariate, we further conduct multiple hypothesis testing by controlling the false discovery rate at the nominal level. Then, we show that the multiple hypothesis testing achieves consistent model selection. Simulation studies and empirical examples are presented to illustrate the finite sample performance and the usefulness of the proposed method, respectively. PMID:28663668

Testing a single regression coefficient in high dimensional linear models.

PubMed

Lan, Wei; Zhong, Ping-Shou; Li, Runze; Wang, Hansheng; Tsai, Chih-Ling

2016-11-01

In linear regression models with high dimensional data, the classical z -test (or t -test) for testing the significance of each single regression coefficient is no longer applicable. This is mainly because the number of covariates exceeds the sample size. In this paper, we propose a simple and novel alternative by introducing the Correlated Predictors Screening (CPS) method to control for predictors that are highly correlated with the target covariate. Accordingly, the classical ordinary least squares approach can be employed to estimate the regression coefficient associated with the target covariate. In addition, we demonstrate that the resulting estimator is consistent and asymptotically normal even if the random errors are heteroscedastic. This enables us to apply the z -test to assess the significance of each covariate. Based on the p -value obtained from testing the significance of each covariate, we further conduct multiple hypothesis testing by controlling the false discovery rate at the nominal level. Then, we show that the multiple hypothesis testing achieves consistent model selection. Simulation studies and empirical examples are presented to illustrate the finite sample performance and the usefulness of the proposed method, respectively.

A quadratic regression modelling on paddy production in the area of Perlis

NASA Astrophysics Data System (ADS)

Goh, Aizat Hanis Annas; Ali, Zalila; Nor, Norlida Mohd; Baharum, Adam; Ahmad, Wan Muhamad Amir W.

2017-08-01

Polynomial regression models are useful in situations in which the relationship between a response variable and predictor variables is curvilinear. Polynomial regression fits the nonlinear relationship into a least squares linear regression model by decomposing the predictor variables into a kth order polynomial. The polynomial order determines the number of inflexions on the curvilinear fitted line. A second order polynomial forms a quadratic expression (parabolic curve) with either a single maximum or minimum, a third order polynomial forms a cubic expression with both a relative maximum and a minimum. This study used paddy data in the area of Perlis to model paddy production based on paddy cultivation characteristics and environmental characteristics. The results indicated that a quadratic regression model best fits the data and paddy production is affected by urea fertilizer application and the interaction between amount of average rainfall and percentage of area defected by pest and disease. Urea fertilizer application has a quadratic effect in the model which indicated that if the number of days of urea fertilizer application increased, paddy production is expected to decrease until it achieved a minimum value and paddy production is expected to increase at higher number of days of urea application. The decrease in paddy production with an increased in rainfall is greater, the higher the percentage of area defected by pest and disease.

Application of third molar development and eruption models in estimating dental age in Malay sub-adults.

PubMed

Mohd Yusof, Mohd Yusmiaidil Putera; Cauwels, Rita; Deschepper, Ellen; Martens, Luc

2015-08-01

The third molar development (TMD) has been widely utilized as one of the radiographic method for dental age estimation. By using the same radiograph of the same individual, third molar eruption (TME) information can be incorporated to the TMD regression model. This study aims to evaluate the performance of dental age estimation in individual method models and the combined model (TMD and TME) based on the classic regressions of multiple linear and principal component analysis. A sample of 705 digital panoramic radiographs of Malay sub-adults aged between 14.1 and 23.8 years was collected. The techniques described by Gleiser and Hunt (modified by Kohler) and Olze were employed to stage the TMD and TME, respectively. The data was divided to develop three respective models based on the two regressions of multiple linear and principal component analysis. The trained models were then validated on the test sample and the accuracy of age prediction was compared between each model. The coefficient of determination (R²) and root mean square error (RMSE) were calculated. In both genders, adjusted R² yielded an increment in the linear regressions of combined model as compared to the individual models. The overall decrease in RMSE was detected in combined model as compared to TMD (0.03-0.06) and TME (0.2-0.8). In principal component regression, low value of adjusted R(2) and high RMSE except in male were exhibited in combined model. Dental age estimation is better predicted using combined model in multiple linear regression models. Copyright © 2015 Elsevier Ltd and Faculty of Forensic and Legal Medicine. All rights reserved.

Modeling thermal sensation in a Mediterranean climate—a comparison of linear and ordinal models

NASA Astrophysics Data System (ADS)

Pantavou, Katerina; Lykoudis, Spyridon

2014-08-01

A simple thermo-physiological model of outdoor thermal sensation adjusted with psychological factors is developed aiming to predict thermal sensation in Mediterranean climates. Microclimatic measurements simultaneously with interviews on personal and psychological conditions were carried out in a square, a street canyon and a coastal location of the greater urban area of Athens, Greece. Multiple linear and ordinal regression were applied in order to estimate thermal sensation making allowance for all the recorded parameters or specific, empirically selected, subsets producing so-called extensive and empirical models, respectively. Meteorological, thermo-physiological and overall models - considering psychological factors as well - were developed. Predictions were improved when personal and psychological factors were taken into account as compared to meteorological models. The model based on ordinal regression reproduced extreme values of thermal sensation vote more adequately than the linear regression one, while the empirical model produced satisfactory results in relation to the extensive model. The effects of adaptation and expectation on thermal sensation vote were introduced in the models by means of the exposure time, season and preference related to air temperature and irradiation. The assessment of thermal sensation could be a useful criterion in decision making regarding public health, outdoor spaces planning and tourism.

Least Squares Procedures.

ERIC Educational Resources Information Center

Hester, Yvette

Least squares methods are sophisticated mathematical curve fitting procedures used in all classical parametric methods. The linear least squares approximation is most often associated with finding the "line of best fit" or the regression line. Since all statistical analyses are correlational and all classical parametric methods are least…

Application of a parameter-estimation technique to modeling the regional aquifer underlying the eastern Snake River plain, Idaho

USGS Publications Warehouse

Garabedian, Stephen P.

1986-01-01

A nonlinear, least-squares regression technique for the estimation of ground-water flow model parameters was applied to the regional aquifer underlying the eastern Snake River Plain, Idaho. The technique uses a computer program to simulate two-dimensional, steady-state ground-water flow. Hydrologic data for the 1980 water year were used to calculate recharge rates, boundary fluxes, and spring discharges. Ground-water use was estimated from irrigated land maps and crop consumptive-use figures. These estimates of ground-water withdrawal, recharge rates, and boundary flux, along with leakance, were used as known values in the model calibration of transmissivity. Leakance values were adjusted between regression solutions by comparing model-calculated to measured spring discharges. In other simulations, recharge and leakance also were calibrated as prior-information regression parameters, which limits the variation of these parameters using a normalized standard error of estimate. Results from a best-fit model indicate a wide areal range in transmissivity from about 0.05 to 44 feet squared per second and in leakance from about 2.2x10 -9 to 6.0 x 10 -8 feet per second per foot. Along with parameter values, model statistics also were calculated, including the coefficient of correlation between calculated and observed head (0.996), the standard error of the estimates for head (40 feet), and the parameter coefficients of variation (about 10-40 percent). Additional boundary flux was added in some areas during calibration to achieve proper fit to ground-water flow directions. Model fit improved significantly when areas that violated model assumptions were removed. It also improved slightly when y-direction (northwest-southeast) transmissivity values were larger than x-direction (northeast-southwest) transmissivity values. The model was most sensitive to changes in recharge, and in some areas, to changes in transmissivity, particularly near the spring discharge area from Milner Dam to King Hill.

A Modified LS+AR Model to Improve the Accuracy of the Short-term Polar Motion Prediction

NASA Astrophysics Data System (ADS)

Wang, Z. W.; Wang, Q. X.; Ding, Y. Q.; Zhang, J. J.; Liu, S. S.

2017-03-01

There are two problems of the LS (Least Squares)+AR (AutoRegressive) model in polar motion forecast: the inner residual value of LS fitting is reasonable, but the residual value of LS extrapolation is poor; and the LS fitting residual sequence is non-linear. It is unsuitable to establish an AR model for the residual sequence to be forecasted, based on the residual sequence before forecast epoch. In this paper, we make solution to those two problems with two steps. First, restrictions are added to the two endpoints of LS fitting data to fix them on the LS fitting curve. Therefore, the fitting values next to the two endpoints are very close to the observation values. Secondly, we select the interpolation residual sequence of an inward LS fitting curve, which has a similar variation trend as the LS extrapolation residual sequence, as the modeling object of AR for the residual forecast. Calculation examples show that this solution can effectively improve the short-term polar motion prediction accuracy by the LS+AR model. In addition, the comparison results of the forecast models of RLS (Robustified Least Squares)+AR, RLS+ARIMA (AutoRegressive Integrated Moving Average), and LS+ANN (Artificial Neural Network) confirm the feasibility and effectiveness of the solution for the polar motion forecast. The results, especially for the polar motion forecast in the 1-10 days, show that the forecast accuracy of the proposed model can reach the world level.

Radioecological modelling of Polonium-210 and Caesium-137 in lichen-reindeer-man and top predators.

PubMed

Persson, Bertil R R; Gjelsvik, Runhild; Holm, Elis

2018-06-01

This work deals with analysis and modelling of the radionuclides 210 Pb and 210 Po in the food-chain lichen-reindeer-man in addition to 210 Po and 137 Cs in top predators. By using the methods of Partial Least Square Regression (PLSR) the atmospheric deposition of 210 Pb and 210 Po is predicted at the sample locations. Dynamic modelling of the activity concentration with differential equations is fitted to the sample data. Reindeer lichen consumption, gastrointestinal absorption, organ distribution and elimination is derived from information in the literature. Dynamic modelling of transfer of 210 Pb and 210 Po to reindeer meat, liver and bone from lichen consumption, fitted well with data from Sweden and Finland from 1966 to 1971. The activity concentration of 210 Pb in the skeleton in man is modelled by using the results of studying the kinetics of lead in skeleton and blood in lead-workers after end of occupational exposure. The result of modelling 210 Pb and 210 Po activity in skeleton matched well with concentrations of 210 Pb and 210 Po in teeth from reindeer-breeders and autopsy bone samples in Finland. The results of 210 Po and 137 Cs in different tissues of wolf, wolverine and lynx previously published, are analysed with multivariate data processing methods such as Principal Component Analysis PCA, and modelled with the method of Projection to Latent Structures, PLS, or Partial Least Square Regression PLSR. Copyright © 2017 Elsevier Ltd. All rights reserved.

Near-infrared Spectroscopy as a Process Analytical Technology Tool for Monitoring the Parching Process of Traditional Chinese Medicine Based on Two Kinds of Chemical Indicators.

PubMed

Li, Kaiyue; Wang, Weiying; Liu, Yanping; Jiang, Su; Huang, Guo; Ye, Liming

2017-01-01

The active ingredients and thus pharmacological efficacy of traditional Chinese medicine (TCM) at different degrees of parching process vary greatly. Near-infrared spectroscopy (NIR) was used to develop a new method for rapid online analysis of TCM parching process, using two kinds of chemical indicators (5-(hydroxymethyl) furfural [5-HMF] content and 420 nm absorbance) as reference values which were obviously observed and changed in most TCM parching process. Three representative TCMs, Areca ( Areca catechu L.), Malt ( Hordeum Vulgare L.), and Hawthorn ( Crataegus pinnatifida Bge.), were used in this study. With partial least squares regression, calibration models of NIR were generated based on two kinds of reference values, i.e. 5-HMF contents measured by high-performance liquid chromatography (HPLC) and 420 nm absorbance measured by ultraviolet-visible spectroscopy (UV/Vis), respectively. In the optimized models for 5-HMF, the root mean square errors of prediction (RMSEP) for Areca, Malt, and Hawthorn was 0.0192, 0.0301, and 0.2600 and correlation coefficients ( R cal ) were 99.86%, 99.88%, and 99.88%, respectively. Moreover, in the optimized models using 420 nm absorbance as reference values, the RMSEP for Areca, Malt, and Hawthorn was 0.0229, 0.0096, and 0.0409 and R cal were 99.69%, 99.81%, and 99.62%, respectively. NIR models with 5-HMF content and 420 nm absorbance as reference values can rapidly and effectively identify three kinds of TCM in different parching processes. This method has great promise to replace current subjective color judgment and time-consuming HPLC or UV/Vis methods and is suitable for rapid online analysis and quality control in TCM industrial manufacturing process. Near-infrared spectroscopy.(NIR) was used to develop a new method for online analysis of traditional Chinese medicine.(TCM) parching processCalibration and validation models of Areca, Malt, and Hawthorn were generated by partial least squares regression using 5.(hydroxymethyl) furfural contents and 420.nm absorbance as reference values, respectively, which were main indicator components during parching process of most TCMThe established NIR models of three TCMs had low root mean square errors of prediction and high correlation coefficientsThe NIR method has great promise for use in TCM industrial manufacturing processes for rapid online analysis and quality control. Abbreviations used: NIR: Near-infrared Spectroscopy; TCM: Traditional Chinese medicine; Areca: Areca catechu L.; Hawthorn: Crataegus pinnatifida Bge.; Malt: Hordeum vulgare L.; 5-HMF: 5-(hydroxymethyl) furfural; PLS: Partial least squares; D: Dimension faction; SLS: Straight line subtraction, MSC: Multiplicative scatter correction; VN: Vector normalization; RMSECV: Root mean square errors of cross-validation; RMSEP: Root mean square errors of validation; R cal : Correlation coefficients; RPD: Residual predictive deviation; PAT: Process analytical technology; FDA: Food and Drug Administration; ICH: International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use.

Reduced Lung Cancer Mortality With Lower Atmospheric Pressure.

PubMed

Merrill, Ray M; Frutos, Aaron

2018-01-01

Research has shown that higher altitude is associated with lower risk of lung cancer and improved survival among patients. The current study assessed the influence of county-level atmospheric pressure (a measure reflecting both altitude and temperature) on age-adjusted lung cancer mortality rates in the contiguous United States, with 2 forms of spatial regression. Ordinary least squares regression and geographically weighted regression models were used to evaluate the impact of climate and other selected variables on lung cancer mortality, based on 2974 counties. Atmospheric pressure was significantly positively associated with lung cancer mortality, after controlling for sunlight, precipitation, PM2.5 (µg/m 3 ), current smoker, and other selected variables. Positive county-level β coefficient estimates ( P < .05) for atmospheric pressure were observed throughout the United States, higher in the eastern half of the country. The spatial regression models showed that atmospheric pressure is positively associated with age-adjusted lung cancer mortality rates, after controlling for other selected variables.

Developing and Testing a Model to Predict Outcomes of Organizational Change

PubMed Central

Gustafson, David H; Sainfort, François; Eichler, Mary; Adams, Laura; Bisognano, Maureen; Steudel, Harold

2003-01-01

Objective To test the effectiveness of a Bayesian model employing subjective probability estimates for predicting success and failure of health care improvement projects. Data Sources Experts' subjective assessment data for model development and independent retrospective data on 221 healthcare improvement projects in the United States, Canada, and the Netherlands collected between 1996 and 2000 for validation. Methods A panel of theoretical and practical experts and literature in organizational change were used to identify factors predicting the outcome of improvement efforts. A Bayesian model was developed to estimate probability of successful change using subjective estimates of likelihood ratios and prior odds elicited from the panel of experts. A subsequent retrospective empirical analysis of change efforts in 198 health care organizations was performed to validate the model. Logistic regression and ROC analysis were used to evaluate the model's performance using three alternative definitions of success. Data Collection For the model development, experts' subjective assessments were elicited using an integrative group process. For the validation study, a staff person intimately involved in each improvement project responded to a written survey asking questions about model factors and project outcomes. Results Logistic regression chi-square statistics and areas under the ROC curve demonstrated a high level of model performance in predicting success. Chi-square statistics were significant at the 0.001 level and areas under the ROC curve were greater than 0.84. Conclusions A subjective Bayesian model was effective in predicting the outcome of actual improvement projects. Additional prospective evaluations as well as testing the impact of this model as an intervention are warranted. PMID:12785571

Confirmatory factor analysis of the female sexual function index.

PubMed

Opperman, Emily A; Benson, Lindsay E; Milhausen, Robin R

2013-01-01

The Female Sexual Functioning Index (Rosen et al., 2000 ) was designed to assess the key dimensions of female sexual functioning using six domains: desire, arousal, lubrication, orgasm, satisfaction, and pain. A full-scale score was proposed to represent women's overall sexual function. The fifth revision to the Diagnostic and Statistical Manual (DSM) is currently underway and includes a proposal to combine desire and arousal problems. The objective of this article was to evaluate and compare four models of the Female Sexual Functioning Index: (a) single-factor model, (b) six-factor model, (c) second-order factor model, and (4) five-factor model combining the desire and arousal subscales. Cross-sectional and observational data from 85 women were used to conduct a confirmatory factor analysis on the Female Sexual Functioning Index. Local and global goodness-of-fit measures, the chi-square test of differences, squared multiple correlations, and regression weights were used. The single-factor model fit was not acceptable. The original six-factor model was confirmed, and good model fit was found for the second-order and five-factor models. Delta chi-square tests of differences supported best fit for the six-factor model validating usage of the six domains. However, when revisions are made to the DSM-5, the Female Sexual Functioning Index can adapt to reflect these changes and remain a valid assessment tool for women's sexual functioning, as the five-factor structure was also supported.

Prediction of p38 map kinase inhibitory activity of 3, 4-dihydropyrido [3, 2-d] pyrimidone derivatives using an expert system based on principal component analysis and least square support vector machine

PubMed Central

Shahlaei, M.; Saghaie, L.

2014-01-01

A quantitative structure–activity relationship (QSAR) study is suggested for the prediction of biological activity (pIC50) of 3, 4-dihydropyrido [3,2-d] pyrimidone derivatives as p38 inhibitors. Modeling of the biological activities of compounds of interest as a function of molecular structures was established by means of principal component analysis (PCA) and least square support vector machine (LS-SVM) methods. The results showed that the pIC50 values calculated by LS-SVM are in good agreement with the experimental data, and the performance of the LS-SVM regression model is superior to the PCA-based model. The developed LS-SVM model was applied for the prediction of the biological activities of pyrimidone derivatives, which were not in the modeling procedure. The resulted model showed high prediction ability with root mean square error of prediction of 0.460 for LS-SVM. The study provided a novel and effective approach for predicting biological activities of 3, 4-dihydropyrido [3,2-d] pyrimidone derivatives as p38 inhibitors and disclosed that LS-SVM can be used as a powerful chemometrics tool for QSAR studies. PMID:26339262

Interpreting the Results of Weighted Least-Squares Regression: Caveats for the Statistical Consumer.

ERIC Educational Resources Information Center

Willett, John B.; Singer, Judith D.

In research, data sets often occur in which the variance of the distribution of the dependent variable at given levels of the predictors is a function of the values of the predictors. In this situation, the use of weighted least-squares (WLS) or techniques is required. Weights suitable for use in a WLS regression analysis must be estimated. A…

Using multiple calibration sets to improve the quantitative accuracy of partial least squares (PLS) regression on open-path fourier transform infrared (OP/FT-IR) spectra of ammonia over wide concentration ranges

USDA-ARS?s Scientific Manuscript database

A technique of using multiple calibration sets in partial least squares regression (PLS) was proposed to improve the quantitative determination of ammonia from open-path Fourier transform infrared spectra. The spectra were measured near animal farms, and the path-integrated concentration of ammonia...

Analysis and improvement measures of flight delay in China

NASA Astrophysics Data System (ADS)

Zang, Yuhang

2017-03-01

Firstly, this paper establishes the principal component regression model to analyze the data quantitatively, based on principal component analysis to get the three principal component factors of flight delays. Then the least square method is used to analyze the factors and obtained the regression equation expression by substitution, and then found that the main reason for flight delays is airlines, followed by weather and traffic. Aiming at the above problems, this paper improves the controllable aspects of traffic flow control. For reasons of traffic flow control, an adaptive genetic queuing model is established for the runway terminal area. This paper, establish optimization method that fifteen planes landed simultaneously on the three runway based on Beijing capital international airport, comparing the results with the existing FCFS algorithm, the superiority of the model is proved.

Statistical model to perform error analysis of curve fits of wind tunnel test data using the techniques of analysis of variance and regression analysis

NASA Technical Reports Server (NTRS)

Alston, D. W.

1981-01-01

The considered research had the objective to design a statistical model that could perform an error analysis of curve fits of wind tunnel test data using analysis of variance and regression analysis techniques. Four related subproblems were defined, and by solving each of these a solution to the general research problem was obtained. The capabilities of the evolved true statistical model are considered. The least squares fit is used to determine the nature of the force, moment, and pressure data. The order of the curve fit is increased in order to delete the quadratic effect in the residuals. The analysis of variance is used to determine the magnitude and effect of the error factor associated with the experimental data.

Local Linear Regression for Data with AR Errors.

PubMed

Li, Runze; Li, Yan

2009-07-01

In many statistical applications, data are collected over time, and they are likely correlated. In this paper, we investigate how to incorporate the correlation information into the local linear regression. Under the assumption that the error process is an auto-regressive process, a new estimation procedure is proposed for the nonparametric regression by using local linear regression method and the profile least squares techniques. We further propose the SCAD penalized profile least squares method to determine the order of auto-regressive process. Extensive Monte Carlo simulation studies are conducted to examine the finite sample performance of the proposed procedure, and to compare the performance of the proposed procedures with the existing one. From our empirical studies, the newly proposed procedures can dramatically improve the accuracy of naive local linear regression with working-independent error structure. We illustrate the proposed methodology by an analysis of real data set.

An augmented classical least squares method for quantitative Raman spectral analysis against component information loss.

PubMed

Zhou, Yan; Cao, Hui

2013-01-01

We propose an augmented classical least squares (ACLS) calibration method for quantitative Raman spectral analysis against component information loss. The Raman spectral signals with low analyte concentration correlations were selected and used as the substitutes for unknown quantitative component information during the CLS calibration procedure. The number of selected signals was determined by using the leave-one-out root-mean-square error of cross-validation (RMSECV) curve. An ACLS model was built based on the augmented concentration matrix and the reference spectral signal matrix. The proposed method was compared with partial least squares (PLS) and principal component regression (PCR) using one example: a data set recorded from an experiment of analyte concentration determination using Raman spectroscopy. A 2-fold cross-validation with Venetian blinds strategy was exploited to evaluate the predictive power of the proposed method. The one-way variance analysis (ANOVA) was used to access the predictive power difference between the proposed method and existing methods. Results indicated that the proposed method is effective at increasing the robust predictive power of traditional CLS model against component information loss and its predictive power is comparable to that of PLS or PCR.

Estuarine Sediment Deposition during Wetland Restoration: A GIS and Remote Sensing Modeling Approach

NASA Technical Reports Server (NTRS)

Newcomer, Michelle; Kuss, Amber; Kentron, Tyler; Remar, Alex; Choksi, Vivek; Skiles, J. W.

2011-01-01

Restoration of the industrial salt flats in the San Francisco Bay, California is an ongoing wetland rehabilitation project. Remote sensing maps of suspended sediment concentration, and other GIS predictor variables were used to model sediment deposition within these recently restored ponds. Suspended sediment concentrations were calibrated to reflectance values from Landsat TM 5 and ASTER using three statistical techniques -- linear regression, multivariate regression, and an Artificial Neural Network (ANN), to map suspended sediment concentrations. Multivariate and ANN regressions using ASTER proved to be the most accurate methods, yielding r2 values of 0.88 and 0.87, respectively. Predictor variables such as sediment grain size and tidal frequency were used in the Marsh Sedimentation (MARSED) model for predicting deposition rates for three years. MARSED results for a fully restored pond show a root mean square deviation (RMSD) of 66.8 mm (<1) between modeled and field observations. This model was further applied to a pond breached in November 2010 and indicated that the recently breached pond will reach equilibrium levels after 60 months of tidal inundation.

Mapping diffuse photosynthetically active radiation from satellite data in Thailand

NASA Astrophysics Data System (ADS)

Choosri, P.; Janjai, S.; Nunez, M.; Buntoung, S.; Charuchittipan, D.

2017-12-01

In this paper, calculation of monthly average hourly diffuse photosynthetically active radiation (PAR) using satellite data is proposed. Diffuse PAR was analyzed at four stations in Thailand. A radiative transfer model was used for calculating the diffuse PAR for cloudless sky conditions. Differences between the diffuse PAR under all sky conditions obtained from the ground-based measurements and those from the model are representative of cloud effects. Two models are developed, one describing diffuse PAR only as a function of solar zenith angle, and the second one as a multiple linear regression with solar zenith angle and satellite reflectivity acting linearly and aerosol optical depth acting in logarithmic functions. When tested with an independent data set, the multiple regression model performed best with a higher coefficient of variance R2 (0.78 vs. 0.70), lower root mean square difference (RMSD) (12.92% vs. 13.05%) and the same mean bias difference (MBD) of -2.20%. Results from the multiple regression model are used to map diffuse PAR throughout the country as monthly averages of hourly data.

Vegetation Fraction Mapping with High Resolution Multispectral Data in the Texas High Plains

NASA Astrophysics Data System (ADS)

Oshaughnessy, S. A.; Gowda, P. H.; Basu, S.; Colaizzi, P. D.; Howell, T. A.; Schulthess, U.

2010-12-01

Land surface models use vegetation fraction to more accurately partition latent, sensible and soil heat fluxes from a partially vegetated surface as it affects energy and moisture exchanges between the earth’s surface and atmosphere. In recent years, there is interest to integrate vegetation fraction data into intelligent irrigation scheduling systems to avoid false positive signals to irrigate. Remote sensing can facilitate the collection of vegetation fraction information on individual fields over large areas in a timely and cost-effective manner. In this study, we developed and evaluated a set of vegetation fraction models using least square regression and artificial neural network (ANN) techniques using RapidEye satellite data (6.5 m spatial resolution and on-demand temporal resolution). Four images were acquired during the 2010 summer growing season, covering bare soil to full crop cover conditions, over the USDA-ARS-Conservation and Production Research Laboratory in Bushland, Texas [350 11' N, 1020 06' W; 1,170 m elevation MSL]. Spectral signatures were extracted from 25 ground truth locations with geographic coordinates. Vegetation fraction information was derived from digital photos taken at the time of image acquisition using a supervised classification technique. Comparison of performance statistics indicate that ANN performed slightly better than least square regression models.

Very-short-term wind power prediction by a hybrid model with single- and multi-step approaches

NASA Astrophysics Data System (ADS)

Mohammed, E.; Wang, S.; Yu, J.

2017-05-01

Very-short-term wind power prediction (VSTWPP) has played an essential role for the operation of electric power systems. This paper aims at improving and applying a hybrid method of VSTWPP based on historical data. The hybrid method is combined by multiple linear regressions and least square (MLR&LS), which is intended for reducing prediction errors. The predicted values are obtained through two sub-processes:1) transform the time-series data of actual wind power into the power ratio, and then predict the power ratio;2) use the predicted power ratio to predict the wind power. Besides, the proposed method can include two prediction approaches: single-step prediction (SSP) and multi-step prediction (MSP). WPP is tested comparatively by auto-regressive moving average (ARMA) model from the predicted values and errors. The validity of the proposed hybrid method is confirmed in terms of error analysis by using probability density function (PDF), mean absolute percent error (MAPE) and means square error (MSE). Meanwhile, comparison of the correlation coefficients between the actual values and the predicted values for different prediction times and window has confirmed that MSP approach by using the hybrid model is the most accurate while comparing to SSP approach and ARMA. The MLR&LS is accurate and promising for solving problems in WPP.

Predicting the composition of red wine blends using an array of multicomponent Peptide-based sensors.

PubMed

Ghanem, Eman; Hopfer, Helene; Navarro, Andrea; Ritzer, Maxwell S; Mahmood, Lina; Fredell, Morgan; Cubley, Ashley; Bolen, Jessica; Fattah, Rabia; Teasdale, Katherine; Lieu, Linh; Chua, Tedmund; Marini, Federico; Heymann, Hildegarde; Anslyn, Eric V

2015-05-20

Differential sensing using synthetic receptors as mimics of the mammalian senses of taste and smell is a powerful approach for the analysis of complex mixtures. Herein, we report on the effectiveness of a cross-reactive, supramolecular, peptide-based sensing array in differentiating and predicting the composition of red wine blends. Fifteen blends of Cabernet Sauvignon, Merlot and Cabernet Franc, in addition to the mono varietals, were used in this investigation. Linear Discriminant Analysis (LDA) showed a clear differentiation of blends based on tannin concentration and composition where certain mono varietals like Cabernet Sauvignon seemed to contribute less to the overall characteristics of the blend. Partial Least Squares (PLS) Regression and cross validation were used to build a predictive model for the responses of the receptors to eleven binary blends and the three mono varietals. The optimized model was later used to predict the percentage of each mono varietal in an independent test set composted of four tri-blends with a 15% average error. A partial least square regression model using the mouth-feel and taste descriptive sensory attributes of the wine blends revealed a strong correlation of the receptors to perceived astringency, which is indicative of selective binding to polyphenols in wine.

Firmness prediction in Prunus persica 'Calrico' peaches by visible/short-wave near infrared spectroscopy and acoustic measurements using optimised linear and non-linear chemometric models.

PubMed

Lafuente, Victoria; Herrera, Luis J; Pérez, María del Mar; Val, Jesús; Negueruela, Ignacio

2015-08-15

In this work, near infrared spectroscopy (NIR) and an acoustic measure (AWETA) (two non-destructive methods) were applied in Prunus persica fruit 'Calrico' (n = 260) to predict Magness-Taylor (MT) firmness. Separate and combined use of these measures was evaluated and compared using partial least squares (PLS) and least squares support vector machine (LS-SVM) regression methods. Also, a mutual-information-based variable selection method, seeking to find the most significant variables to produce optimal accuracy of the regression models, was applied to a joint set of variables (NIR wavelengths and AWETA measure). The newly proposed combined NIR-AWETA model gave good values of the determination coefficient (R(2)) for PLS and LS-SVM methods (0.77 and 0.78, respectively), improving the reliability of MT firmness prediction in comparison with separate NIR and AWETA predictions. The three variables selected by the variable selection method (AWETA measure plus NIR wavelengths 675 and 697 nm) achieved R(2) values 0.76 and 0.77, PLS and LS-SVM. These results indicated that the proposed mutual-information-based variable selection algorithm was a powerful tool for the selection of the most relevant variables. © 2014 Society of Chemical Industry.

Determination of total phenolic compounds in compost by infrared spectroscopy.

PubMed

Cascant, M M; Sisouane, M; Tahiri, S; Krati, M El; Cervera, M L; Garrigues, S; de la Guardia, M

2016-06-01

Middle and near infrared (MIR and NIR) were applied to determine the total phenolic compounds (TPC) content in compost samples based on models built by using partial least squares (PLS) regression. The multiplicative scatter correction, standard normal variate and first derivative were employed as spectra pretreatment, and the number of latent variable were optimized by leave-one-out cross-validation. The performance of PLS-ATR-MIR and PLS-DR-NIR models was evaluated according to root mean square error of cross validation and prediction (RMSECV and RMSEP), the coefficient of determination for prediction (Rpred(2)) and residual predictive deviation (RPD) being obtained for this latter values of 5.83 and 8.26 for MIR and NIR, respectively. Copyright © 2016 Elsevier B.V. All rights reserved.

Application of Fourier transform near-infrared spectroscopy combined with high-performance liquid chromatography in rapid and simultaneous determination of essential components in crude Radix Scrophulariae.

PubMed

Li, Xiaomeng; Fang, Dansi; Cong, Xiaodong; Cao, Gang; Cai, Hao; Cai, Baochang

2012-12-01

A method is described using rapid and sensitive Fourier transform near-infrared spectroscopy combined with high-performance liquid chromatography-diode array detection for the simultaneous identification and determination of four bioactive compounds in crude Radix Scrophulariae samples. Partial least squares regression is selected as the analysis type and multiplicative scatter correction, second derivative, and Savitzky-Golay filter were adopted for the spectral pretreatment. The correlation coefficients (R) of the calibration models were above 0.96 and the root mean square error of predictions were under 0.028. The developed models were applied to unknown samples with satisfactory results. The established method was validated and can be applied to the intrinsic quality control of crude Radix Scrophulariae.

Multivariate methods on the excitation emission matrix fluorescence spectroscopic data of diesel-kerosene mixtures: a comparative study.

PubMed

Divya, O; Mishra, Ashok K

2007-05-29

Quantitative determination of kerosene fraction present in diesel has been carried out based on excitation emission matrix fluorescence (EEMF) along with parallel factor analysis (PARAFAC) and N-way partial least squares regression (N-PLS). EEMF is a simple, sensitive and nondestructive method suitable for the analysis of multifluorophoric mixtures. Calibration models consisting of varying compositions of diesel and kerosene were constructed and their validation was carried out using leave-one-out cross validation method. The accuracy of the model was evaluated through the root mean square error of prediction (RMSEP) for the PARAFAC, N-PLS and unfold PLS methods. N-PLS was found to be a better method compared to PARAFAC and unfold PLS method because of its low RMSEP values.

On structure-exploiting trust-region regularized nonlinear least squares algorithms for neural-network learning.

PubMed

Mizutani, Eiji; Demmel, James W

2003-01-01

This paper briefly introduces our numerical linear algebra approaches for solving structured nonlinear least squares problems arising from 'multiple-output' neural-network (NN) models. Our algorithms feature trust-region regularization, and exploit sparsity of either the 'block-angular' residual Jacobian matrix or the 'block-arrow' Gauss-Newton Hessian (or Fisher information matrix in statistical sense) depending on problem scale so as to render a large class of NN-learning algorithms 'efficient' in both memory and operation costs. Using a relatively large real-world nonlinear regression application, we shall explain algorithmic strengths and weaknesses, analyzing simulation results obtained by both direct and iterative trust-region algorithms with two distinct NN models: 'multilayer perceptrons' (MLP) and 'complementary mixtures of MLP-experts' (or neuro-fuzzy modular networks).

A robust ridge regression approach in the presence of both multicollinearity and outliers in the data

NASA Astrophysics Data System (ADS)

Shariff, Nurul Sima Mohamad; Ferdaos, Nur Aqilah

2017-08-01

Multicollinearity often leads to inconsistent and unreliable parameter estimates in regression analysis. This situation will be more severe in the presence of outliers it will cause fatter tails in the error distributions than the normal distributions. The well-known procedure that is robust to multicollinearity problem is the ridge regression method. This method however is expected to be affected by the presence of outliers due to some assumptions imposed in the modeling procedure. Thus, the robust version of existing ridge method with some modification in the inverse matrix and the estimated response value is introduced. The performance of the proposed method is discussed and comparisons are made with several existing estimators namely, Ordinary Least Squares (OLS), ridge regression and robust ridge regression based on GM-estimates. The finding of this study is able to produce reliable parameter estimates in the presence of both multicollinearity and outliers in the data.

Methods for estimating magnitude and frequency of 1-, 3-, 7-, 15-, and 30-day flood-duration flows in Arizona

USGS Publications Warehouse

Kennedy, Jeffrey R.; Paretti, Nicholas V.; Veilleux, Andrea G.

2014-01-01

Regression equations, which allow predictions of n-day flood-duration flows for selected annual exceedance probabilities at ungaged sites, were developed using generalized least-squares regression and flood-duration flow frequency estimates at 56 streamgaging stations within a single, relatively uniform physiographic region in the central part of Arizona, between the Colorado Plateau and Basin and Range Province, called the Transition Zone. Drainage area explained most of the variation in the n-day flood-duration annual exceedance probabilities, but mean annual precipitation and mean elevation were also significant variables in the regression models. Standard error of prediction for the regression equations varies from 28 to 53 percent and generally decreases with increasing n-day duration. Outside the Transition Zone there are insufficient streamgaging stations to develop regression equations, but flood-duration flow frequency estimates are presented at select streamgaging stations.

Order-constrained linear optimization.

PubMed

Tidwell, Joe W; Dougherty, Michael R; Chrabaszcz, Jeffrey S; Thomas, Rick P

2017-11-01

Despite the fact that data and theories in the social, behavioural, and health sciences are often represented on an ordinal scale, there has been relatively little emphasis on modelling ordinal properties. The most common analytic framework used in psychological science is the general linear model, whose variants include ANOVA, MANOVA, and ordinary linear regression. While these methods are designed to provide the best fit to the metric properties of the data, they are not designed to maximally model ordinal properties. In this paper, we develop an order-constrained linear least-squares (OCLO) optimization algorithm that maximizes the linear least-squares fit to the data conditional on maximizing the ordinal fit based on Kendall's τ. The algorithm builds on the maximum rank correlation estimator (Han, 1987, Journal of Econometrics, 35, 303) and the general monotone model (Dougherty & Thomas, 2012, Psychological Review, 119, 321). Analyses of simulated data indicate that when modelling data that adhere to the assumptions of ordinary least squares, OCLO shows minimal bias, little increase in variance, and almost no loss in out-of-sample predictive accuracy. In contrast, under conditions in which data include a small number of extreme scores (fat-tailed distributions), OCLO shows less bias and variance, and substantially better out-of-sample predictive accuracy, even when the outliers are removed. We show that the advantages of OCLO over ordinary least squares in predicting new observations hold across a variety of scenarios in which researchers must decide to retain or eliminate extreme scores when fitting data. © 2017 The British Psychological Society.

Estimation of peak-discharge frequency of urban streams in Jefferson County, Kentucky

USGS Publications Warehouse

Martin, Gary R.; Ruhl, Kevin J.; Moore, Brian L.; Rose, Martin F.

1997-01-01

An investigation of flood-hydrograph characteristics for streams in urban Jefferson County, Kentucky, was made to obtain hydrologic information needed for waterresources management. Equations for estimating peak-discharge frequencies for ungaged streams in the county were developed by combining (1) long-term annual peakdischarge data and rainfall-runoff data collected from 1991 to 1995 in 13 urban basins and (2) long-term annual peak-discharge data in four rural basins located in hydrologically similar areas of neighboring counties. The basins ranged in size from 1.36 to 64.0 square miles. The U.S. Geological Survey Rainfall- Runoff Model (RRM) was calibrated for each of the urban basins. The calibrated models were used with long-term, historical rainfall and pan-evaporation data to simulate 79 years of annual peak-discharge data. Peak-discharge frequencies were estimated by fitting the logarithms of the annual peak discharges to a Pearson-Type III frequency distribution. The simulated peak-discharge frequencies were adjusted for improved reliability by application of bias-correction factors derived from peakdischarge frequencies based on local, observed annual peak discharges. The three-parameter and the preferred seven-parameter nationwide urban-peak-discharge regression equations previously developed by USGS investigators provided biased (high) estimates for the urban basins studied. Generalized-least-square regression procedures were used to relate peakdischarge frequency to selected basin characteristics. Regression equations were developed to estimate peak-discharge frequency by adjusting peak-dischargefrequency estimates made by use of the threeparameter nationwide urban regression equations. The regression equations are presented in equivalent forms as functions of contributing drainage area, main-channel slope, and basin development factor, which is an index for measuring the efficiency of the basin drainage system. Estimates of peak discharges for streams in the county can be made for the 2-, 5-, 10-, 25-, 50-, and 100-year recurrence intervals by use of the regression equations. The average standard errors of prediction of the regression equations ranges from ? 34 to ? 45 percent. The regression equations are applicable to ungaged streams in the county having a specific range of basin characteristics.

Generalized additive regression models of discharge and mean velocity associated with direct-runoff conditions in Texas: Utility of the U.S. Geological Survey discharge measurement database

USGS Publications Warehouse

Asquith, William H.; Herrmann, George R.; Cleveland, Theodore G.

2013-01-01

A database containing more than 17,700 discharge values and ancillary hydraulic properties was assembled from summaries of discharge measurement records for 424 U.S. Geological Survey streamflow-gauging stations (stream gauges) in Texas. Each discharge exceeds the 90th-percentile daily mean streamflow as determined by period-of-record, stream-gauge-specific, flow-duration curves. Each discharge therefore is assumed to represent discharge measurement made during direct-runoff conditions. The hydraulic properties of each discharge measurement included concomitant cross-sectional flow area, water-surface top width, and reported mean velocity. Systematic and statewide investigation of these data in pursuit of regional models for the estimation of discharge and mean velocity has not been previously attempted. Generalized additive regression modeling is used to develop readily implemented procedures by end-users for estimation of discharge and mean velocity from select predictor variables at ungauged stream locations. The discharge model uses predictor variables of cross-sectional flow area, top width, stream location, mean annual precipitation, and a generalized terrain and climate index (OmegaEM) derived for a previous flood-frequency regionalization study. The mean velocity model uses predictor variables of discharge, top width, stream location, mean annual precipitation, and OmegaEM. The discharge model has an adjusted R-squared value of about 0.95 and a residual standard error (RSE) of about 0.22 base-10 logarithm (cubic meters per second); the mean velocity model has an adjusted R-squared value of about 0.67 and an RSE of about 0.063 fifth root (meters per second). Example applications and computations using both regression models are provided. - See more at: http://ascelibrary.org/doi/abs/10.1061/%28ASCE%29HE.1943-5584.0000635#sthash.jhGyPxgZ.dpuf

Robust ridge regression estimators for nonlinear models with applications to high throughput screening assay data.

PubMed

Lim, Changwon

2015-03-30

Nonlinear regression is often used to evaluate the toxicity of a chemical or a drug by fitting data from a dose-response study. Toxicologists and pharmacologists may draw a conclusion about whether a chemical is toxic by testing the significance of the estimated parameters. However, sometimes the null hypothesis cannot be rejected even though the fit is quite good. One possible reason for such cases is that the estimated standard errors of the parameter estimates are extremely large. In this paper, we propose robust ridge regression estimation procedures for nonlinear models to solve this problem. The asymptotic properties of the proposed estimators are investigated; in particular, their mean squared errors are derived. The performances of the proposed estimators are compared with several standard estimators using simulation studies. The proposed methodology is also illustrated using high throughput screening assay data obtained from the National Toxicology Program. Copyright © 2014 John Wiley & Sons, Ltd.

Artificial bias typically neglected in comparisons of uncertain atmospheric data

NASA Astrophysics Data System (ADS)

Pitkänen, Mikko R. A.; Mikkonen, Santtu; Lehtinen, Kari E. J.; Lipponen, Antti; Arola, Antti

2016-09-01

Publications in atmospheric sciences typically neglect biases caused by regression dilution (bias of the ordinary least squares line fitting) and regression to the mean (RTM) in comparisons of uncertain data. We use synthetic observations mimicking real atmospheric data to demonstrate how the biases arise from random data uncertainties of measurements, model output, or satellite retrieval products. Further, we provide examples of typical methods of data comparisons that have a tendency to pronounce the biases. The results show, that data uncertainties can significantly bias data comparisons due to regression dilution and RTM, a fact that is known in statistics but disregarded in atmospheric sciences. Thus, we argue that often these biases are widely regarded as measurement or modeling errors, for instance, while they in fact are artificial. It is essential that atmospheric and geoscience communities become aware of and consider these features in research.

Prediction of Patient-Controlled Analgesic Consumption: A Multimodel Regression Tree Approach.

PubMed

Hu, Yuh-Jyh; Ku, Tien-Hsiung; Yang, Yu-Hung; Shen, Jia-Ying

2018-01-01

Several factors contribute to individual variability in postoperative pain, therefore, individuals consume postoperative analgesics at different rates. Although many statistical studies have analyzed postoperative pain and analgesic consumption, most have identified only the correlation and have not subjected the statistical model to further tests in order to evaluate its predictive accuracy. In this study involving 3052 patients, a multistrategy computational approach was developed for analgesic consumption prediction. This approach uses data on patient-controlled analgesia demand behavior over time and combines clustering, classification, and regression to mitigate the limitations of current statistical models. Cross-validation results indicated that the proposed approach significantly outperforms various existing regression methods. Moreover, a comparison between the predictions by anesthesiologists and medical specialists and those of the computational approach for an independent test data set of 60 patients further evidenced the superiority of the computational approach in predicting analgesic consumption because it produced markedly lower root mean squared errors.

FPGA implementation of predictive degradation model for engine oil lifetime

NASA Astrophysics Data System (ADS)

Idros, M. F. M.; Razak, A. H. A.; Junid, S. A. M. Al; Suliman, S. I.; Halim, A. K.

2018-03-01

This paper presents the implementation of linear regression model for degradation prediction on Register Transfer Logic (RTL) using QuartusII. A stationary model had been identified in the degradation trend for the engine oil in a vehicle in time series method. As for RTL implementation, the degradation model is written in Verilog HDL and the data input are taken at a certain time. Clock divider had been designed to support the timing sequence of input data. At every five data, a regression analysis is adapted for slope variation determination and prediction calculation. Here, only the negative value are taken as the consideration for the prediction purposes for less number of logic gate. Least Square Method is adapted to get the best linear model based on the mean values of time series data. The coded algorithm has been implemented on FPGA for validation purposes. The result shows the prediction time to change the engine oil.

Predicting seasonal influenza transmission using functional regression models with temporal dependence.

PubMed

Oviedo de la Fuente, Manuel; Febrero-Bande, Manuel; Muñoz, María Pilar; Domínguez, Àngela

2018-01-01

This paper proposes a novel approach that uses meteorological information to predict the incidence of influenza in Galicia (Spain). It extends the Generalized Least Squares (GLS) methods in the multivariate framework to functional regression models with dependent errors. These kinds of models are useful when the recent history of the incidence of influenza are readily unavailable (for instance, by delays on the communication with health informants) and the prediction must be constructed by correcting the temporal dependence of the residuals and using more accessible variables. A simulation study shows that the GLS estimators render better estimations of the parameters associated with the regression model than they do with the classical models. They obtain extremely good results from the predictive point of view and are competitive with the classical time series approach for the incidence of influenza. An iterative version of the GLS estimator (called iGLS) was also proposed that can help to model complicated dependence structures. For constructing the model, the distance correlation measure [Formula: see text] was employed to select relevant information to predict influenza rate mixing multivariate and functional variables. These kinds of models are extremely useful to health managers in allocating resources in advance to manage influenza epidemics.

Application of a Combined Model with Autoregressive Integrated Moving Average (ARIMA) and Generalized Regression Neural Network (GRNN) in Forecasting Hepatitis Incidence in Heng County, China

PubMed Central

Liang, Hao; Gao, Lian; Liang, Bingyu; Huang, Jiegang; Zang, Ning; Liao, Yanyan; Yu, Jun; Lai, Jingzhen; Qin, Fengxiang; Su, Jinming; Ye, Li; Chen, Hui

2016-01-01

Background Hepatitis is a serious public health problem with increasing cases and property damage in Heng County. It is necessary to develop a model to predict the hepatitis epidemic that could be useful for preventing this disease. Methods The autoregressive integrated moving average (ARIMA) model and the generalized regression neural network (GRNN) model were used to fit the incidence data from the Heng County CDC (Center for Disease Control and Prevention) from January 2005 to December 2012. Then, the ARIMA-GRNN hybrid model was developed. The incidence data from January 2013 to December 2013 were used to validate the models. Several parameters, including mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE) and mean square error (MSE), were used to compare the performance among the three models. Results The morbidity of hepatitis from Jan 2005 to Dec 2012 has seasonal variation and slightly rising trend. The ARIMA(0,1,2)(1,1,1)12 model was the most appropriate one with the residual test showing a white noise sequence. The smoothing factor of the basic GRNN model and the combined model was 1.8 and 0.07, respectively. The four parameters of the hybrid model were lower than those of the two single models in the validation. The parameters values of the GRNN model were the lowest in the fitting of the three models. Conclusions The hybrid ARIMA-GRNN model showed better hepatitis incidence forecasting in Heng County than the single ARIMA model and the basic GRNN model. It is a potential decision-supportive tool for controlling hepatitis in Heng County. PMID:27258555

Direct quantification of test bacteria in synthetic water-polluted samples by square wave voltammetry and chemometric methods.

PubMed

Carpani, Irene; Conti, Paolo; Lanteri, Silvia; Legnani, Pier Paolo; Leoni, Erica; Tonelli, Domenica

2008-02-28

A home-made microelectrode array, based on reticulated vitreous carbon, was used as working electrode in square wave voltammetry experiments to quantify the bacterial load of Escherichia coli ATCC 13706 and Pseudomonas aeruginosa ATCC 27853, chosen as test microorganisms, in synthetic samples similar to drinking water (phosphate buffer). Raw electrochemical signals were analysed with partial least squares regression coupled to variable selection in order to correlate these values with the bacterial load estimated by aerobic plate counting. The results demonstrated the ability of the method to detect even low loads of microorganisms in synthetic water samples. In particular, the model detects the bacterial load in the range 3-2,020 CFU ml(-1) for E. coli and in the range 76-155,556 CFU ml(-1) for P. aeruginosa.

Random sample consensus combined with partial least squares regression (RANSAC-PLS) for microbial metabolomics data mining and phenotype improvement.

PubMed

Teoh, Shao Thing; Kitamura, Miki; Nakayama, Yasumune; Putri, Sastia; Mukai, Yukio; Fukusaki, Eiichiro

2016-08-01

In recent years, the advent of high-throughput omics technology has made possible a new class of strain engineering approaches, based on identification of possible gene targets for phenotype improvement from omic-level comparison of different strains or growth conditions. Metabolomics, with its focus on the omic level closest to the phenotype, lends itself naturally to this semi-rational methodology. When a quantitative phenotype such as growth rate under stress is considered, regression modeling using multivariate techniques such as partial least squares (PLS) is often used to identify metabolites correlated with the target phenotype. However, linear modeling techniques such as PLS require a consistent metabolite-phenotype trend across the samples, which may not be the case when outliers or multiple conflicting trends are present in the data. To address this, we proposed a data-mining strategy that utilizes random sample consensus (RANSAC) to select subsets of samples with consistent trends for construction of better regression models. By applying a combination of RANSAC and PLS (RANSAC-PLS) to a dataset from a previous study (gas chromatography/mass spectrometry metabolomics data and 1-butanol tolerance of 19 yeast mutant strains), new metabolites were indicated to be correlated with tolerance within certain subsets of the samples. The relevance of these metabolites to 1-butanol tolerance were then validated from single-deletion strains of corresponding metabolic genes. The results showed that RANSAC-PLS is a promising strategy to identify unique metabolites that provide additional hints for phenotype improvement, which could not be detected by traditional PLS modeling using the entire dataset. Copyright © 2016 The Society for Biotechnology, Japan. Published by Elsevier B.V. All rights reserved.

An automated ranking platform for machine learning regression models for meat spoilage prediction using multi-spectral imaging and metabolic profiling.

PubMed

Estelles-Lopez, Lucia; Ropodi, Athina; Pavlidis, Dimitris; Fotopoulou, Jenny; Gkousari, Christina; Peyrodie, Audrey; Panagou, Efstathios; Nychas, George-John; Mohareb, Fady

2017-09-01

Over the past decade, analytical approaches based on vibrational spectroscopy, hyperspectral/multispectral imagining and biomimetic sensors started gaining popularity as rapid and efficient methods for assessing food quality, safety and authentication; as a sensible alternative to the expensive and time-consuming conventional microbiological techniques. Due to the multi-dimensional nature of the data generated from such analyses, the output needs to be coupled with a suitable statistical approach or machine-learning algorithms before the results can be interpreted. Choosing the optimum pattern recognition or machine learning approach for a given analytical platform is often challenging and involves a comparative analysis between various algorithms in order to achieve the best possible prediction accuracy. In this work, "MeatReg", a web-based application is presented, able to automate the procedure of identifying the best machine learning method for comparing data from several analytical techniques, to predict the counts of microorganisms responsible of meat spoilage regardless of the packaging system applied. In particularly up to 7 regression methods were applied and these are ordinary least squares regression, stepwise linear regression, partial least square regression, principal component regression, support vector regression, random forest and k-nearest neighbours. MeatReg" was tested with minced beef samples stored under aerobic and modified atmosphere packaging and analysed with electronic nose, HPLC, FT-IR, GC-MS and Multispectral imaging instrument. Population of total viable count, lactic acid bacteria, pseudomonads, Enterobacteriaceae and B. thermosphacta, were predicted. As a result, recommendations of which analytical platforms are suitable to predict each type of bacteria and which machine learning methods to use in each case were obtained. The developed system is accessible via the link: www.sorfml.com. Copyright © 2017 Elsevier Ltd. All rights reserved.

Recurrence risk model for esophageal cancer after radical surgery.

PubMed

Lu, Jincheng; Tao, Hua; Song, Dan; Chen, Cheng

2013-10-01

The aim of the present study was to construct a risk assessment model which was tested by disease-free survival (DFS) of esophageal cancer after radical surgery. A total of 164 consecutive esophageal cancer patients who had undergone radical surgery between January 2005 and December 2006 were retrospectively analyzed. The cutpoint of value at risk (VaR) was inferred by stem-and-leaf plot, as well as by independent-samples t-test for recurrence-free time, further confirmed by crosstab chi-square test, univariate analysis and Cox regression analysis for DFS. The cutpoint of VaR was 0.3 on the basis of our model. The rate of recurrence was 30.3% (30/99) and 52.3% (34/65) in VaR <0.3 and VaR ≥0.3 (chi-square test, (χ) (2) =7.984, P=0.005), respectively. The 1-, 3-, and 5-year DFS of esophageal cancer after radical surgery was 70.4%, 48.7%, and 45.3%, respectively in VaR ≥0.3, whereas 91.5%, 75.8%, and 67.3%, respectively in VaR <0.3 (Log-rank test, (χ) (2) =9.59, P=0.0020), and further confirmed by Cox regression analysis [hazard ratio =2.10, 95% confidence interval (CI): 1.2649-3.4751; P=0.0041]. The model could be applied for integrated assessment of recurrence risk after radical surgery for esophageal cancer.

Recurrence risk model for esophageal cancer after radical surgery

PubMed Central

Tao, Hua; Song, Dan; Chen, Cheng

2013-01-01

Objective The aim of the present study was to construct a risk assessment model which was tested by disease-free survival (DFS) of esophageal cancer after radical surgery. Methods A total of 164 consecutive esophageal cancer patients who had undergone radical surgery between January 2005 and December 2006 were retrospectively analyzed. The cutpoint of value at risk (VaR) was inferred by stem-and-leaf plot, as well as by independent-samples t-test for recurrence-free time, further confirmed by crosstab chi-square test, univariate analysis and Cox regression analysis for DFS. Results The cutpoint of VaR was 0.3 on the basis of our model. The rate of recurrence was 30.3% (30/99) and 52.3% (34/65) in VaR <0.3 and VaR ≥0.3 (chi-square test, χ2 =7.984, P=0.005), respectively. The 1-, 3-, and 5-year DFS of esophageal cancer after radical surgery was 70.4%, 48.7%, and 45.3%, respectively in VaR ≥0.3, whereas 91.5%, 75.8%, and 67.3%, respectively in VaR <0.3 (Log-rank test, χ2 =9.59, P=0.0020), and further confirmed by Cox regression analysis [hazard ratio =2.10, 95% confidence interval (CI): 1.2649-3.4751; P=0.0041]. Conclusions The model could be applied for integrated assessment of recurrence risk after radical surgery for esophageal cancer. PMID:24255579

Faculty Composition in Four-Year Institutions: The Role of Pressures, Values, and Organizational Processes in Academic Decision-Making

ERIC Educational Resources Information Center

Kezar, Adrianna; Gehrke, Sean

2016-01-01

This study broadens our understanding of conditions that shape faculty composition in higher education. We surveyed academic deans to evaluate their views on the professoriate, values, pressures, and practices pertaining to the use of non-tenure-track faculty (NTTF). We utilized [ordinary-least-squares] OLS regression to test a model for…

Development of the first sphingomyelin biomimetic stationary phase for immobilized artificial membrane (IAM) chromatography.

PubMed

Verzele, Dieter; Lynen, Frédéric; De Vrieze, Mike; Wright, Adrian G; Hanna-Brown, Melissa; Sandra, Pat

2012-01-28

A prototype sphingomyelin stationary phase for Immobilized Artificial Membrane (IAM) chromatography was synthesized by an ultra-short, solid-phase inspired methodology, in which an oxidative release monitoring strategy played a vital role. Evaluated in a proof-of-concept model for blood-brain barrier passage, partial least squares regression demonstrated its potential as an in vitro prediction tool.

On Improved Least Squares Regression and Artificial Neural Network Meta-Models for Simulation via Control Variates

DTIC Science & Technology

2016-09-15

18] under the context of robust parameter design for simulation. Bellucci’s technique is used in this research, primarily because the interior -point...Fundamentals of Radial Basis Neural Network (RBNN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 1.2.2.2 Design of Experiments...with Neural Nets . . . . . . . . . . . . . 31 1.2.2.3 Factorial Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 1.2.2.4

Equations for Estimating Biomass of Herbaceous and Woody Vegetation in Early-Successional Southern Appalachian Pine-Hardwood Forests

Treesearch

Katherine J. Elliott; Barton D. Clinton

1993-01-01

Allometric equations were developed to predict aboveground dry weight of herbaceous and woody species on prescribe-burned sites in the Southern Appalachians. Best-fit least-square regression models were developed using diamet,er, height, or both, as the independent variables and dry weight as the dependent variable. Coefficients of determination for the selected total...

The Effect of Executive Order 13269 on Noncitizen Enlisted Accessions in the U.S. Military

DTIC Science & Technology

2013-03-01

heteroscedasticity by using both White and Breusch - Pagan Tests (Woolridge, 2009). The test results showed that there was significant heteroscedasticy in...39  C.  JOINT-HYPOTHESIS TEST AND SIGNIFICANCE OF THE MODELS...Accessions Regression Results ........................... 40  Table 8.  The Authors’ R Square and F Test Results of DMDC Data ................ 42  xii

A Hedonic Approach to Estimating Software Cost Using Ordinary Least Squares Regression and Nominal Attribute Variables

DTIC Science & Technology

2006-03-01

included zero, there is insufficient evidence to indicate that the error mean is 35 not zero. The Breusch - Pagan test was used to test the constant...Multicollinearity .............................................................................. 33 Testing OLS Assumptions...programming styles used by developers (Stamelos and others, 2003:733). Kemerer tested to see how models utilizing SLOC as an independent variable

Non-destructive prediction of low levels of melamine particles in milk powder using hyperspectral reflectance imaging and partial least square regression model

USDA-ARS?s Scientific Manuscript database

Melamine has been used in industrial manufacturing of numerous forms of plastics, fertilizer, adhesives and laminates. In 2008, dairy products tainted with melamine have been reported to be responsible for kidney stones and renal failure among infants and children in China. Some dairy farmers and ma...

Curriculum-Based Measurement of Reading: An Evaluation of Frequentist and Bayesian Methods to Model Progress Monitoring Data

ERIC Educational Resources Information Center

Christ, Theodore J.; Desjardins, Christopher David

2018-01-01

Curriculum-Based Measurement of Oral Reading (CBM-R) is often used to monitor student progress and guide educational decisions. Ordinary least squares regression (OLSR) is the most widely used method to estimate the slope, or rate of improvement (ROI), even though published research demonstrates OLSR's lack of validity and reliability, and…

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.

Predictive sparse modeling of fMRI data for improved classification, regression, and visualization using the k-support norm.

PubMed

Belilovsky, Eugene; Gkirtzou, Katerina; Misyrlis, Michail; Konova, Anna B; Honorio, Jean; Alia-Klein, Nelly; Goldstein, Rita Z; Samaras, Dimitris; Blaschko, Matthew B

2015-12-01

We explore various sparse regularization techniques for analyzing fMRI data, such as the ℓ1 norm (often called LASSO in the context of a squared loss function), elastic net, and the recently introduced k-support norm. Employing sparsity regularization allows us to handle the curse of dimensionality, a problem commonly found in fMRI analysis. In this work we consider sparse regularization in both the regression and classification settings. We perform experiments on fMRI scans from cocaine-addicted as well as healthy control subjects. We show that in many cases, use of the k-support norm leads to better predictive performance, solution stability, and interpretability as compared to other standard approaches. We additionally analyze the advantages of using the absolute loss function versus the standard squared loss which leads to significantly better predictive performance for the regularization methods tested in almost all cases. Our results support the use of the k-support norm for fMRI analysis and on the clinical side, the generalizability of the I-RISA model of cocaine addiction. Copyright © 2015 Elsevier Ltd. All rights reserved.

The dynamic relationships between economic status and health measures among working-age adults in the United States.

PubMed

Meraya, Abdulkarim M; Dwibedi, Nilanjana; Tan, Xi; Innes, Kim; Mitra, Sophie; Sambamoorthi, Usha

2018-04-18

We examine the dynamic relationships between economic status and health measures using data from 8 waves of the Panel Study of Income Dynamics from 1999 to 2013. Health measures are self-rated health (SRH) and functional limitations; economic status measures are labor income (earnings), family income, and net wealth. We use 3 different types of models: (a) ordinary least squares regression, (b) first-difference, and (c) system-generalized method of moment (GMM). Using ordinary least squares regression and first difference models, we find that higher levels of economic status are associated with better SRH and functional status among both men and women, although declines in income and wealth are associated with a decline in health for men only. Using system-GMM estimators, we find evidence of a causal link from labor income to SRH and functional status for both genders. Among men only, system-GMM results indicate that there is a causal link from net wealth to SRH and functional status. Results overall highlight the need for integrated economic and health policies, and for policies that mitigate the potential adverse health effects of short-term changes in economic status. Copyright © 2018 John Wiley & Sons, Ltd.

A generalized least squares regression approach for computing effect sizes in single-case research: application examples.

PubMed

Maggin, Daniel M; Swaminathan, Hariharan; Rogers, Helen J; O'Keeffe, Breda V; Sugai, George; Horner, Robert H

2011-06-01

A new method for deriving effect sizes from single-case designs is proposed. The strategy is applicable to small-sample time-series data with autoregressive errors. The method uses Generalized Least Squares (GLS) to model the autocorrelation of the data and estimate regression parameters to produce an effect size that represents the magnitude of treatment effect from baseline to treatment phases in standard deviation units. In this paper, the method is applied to two published examples using common single case designs (i.e., withdrawal and multiple-baseline). The results from these studies are described, and the method is compared to ten desirable criteria for single-case effect sizes. Based on the results of this application, we conclude with observations about the use of GLS as a support to visual analysis, provide recommendations for future research, and describe implications for practice. Copyright © 2011 Society for the Study of School Psychology. Published by Elsevier Ltd. All rights reserved.

Development of predictive models for total phenolics and free p-coumaric acid contents in barley grain by near-infrared spectroscopy.

PubMed

Han, Zhigang; Cai, Shengguan; Zhang, Xuelei; Qian, Qiufeng; Huang, Yuqing; Dai, Fei; Zhang, Guoping

2017-07-15

Barley grains are rich in phenolic compounds, which are associated with reduced risk of chronic diseases. Development of barley cultivars with high phenolic acid content has become one of the main objectives in breeding programs. A rapid and accurate method for measuring phenolic compounds would be helpful for crop breeding. We developed predictive models for both total phenolics (TPC) and p-coumaric acid (PA), based on near-infrared spectroscopy (NIRS) analysis. Regressions of partial least squares (PLS) and least squares support vector machine (LS-SVM) were compared for improving the models, and Monte Carlo-Uninformative Variable Elimination (MC-UVE) was applied to select informative wavelengths. The optimal calibration models generated high coefficients of correlation (r pre ) and ratio performance deviation (RPD) for TPC and PA. These results indicated the models are suitable for rapid determination of phenolic compounds in barley grains. Copyright © 2017 Elsevier Ltd. All rights reserved.

An In Silico Method for Screening Nicotine Derivatives as Cytochrome P450 2A6 Selective Inhibitors Based on Kernel Partial Least Squares

PubMed Central

Wang, Yonghua; Li, Yan; Wang, Bin

2007-01-01

Nicotine and a variety of other drugs and toxins are metabolized by cytochrome P450 (CYP) 2A6. The aim of the present study was to build a quantitative structure-activity relationship (QSAR) model to predict the activities of nicotine analogues on CYP2A6. Kernel partial least squares (K-PLS) regression was employed with the electro-topological descriptors to build the computational models. Both the internal and external predictabilities of the models were evaluated with test sets to ensure their validity and reliability. As a comparison to K-PLS, a standard PLS algorithm was also applied on the same training and test sets. Our results show that the K-PLS produced reasonable results that outperformed the PLS model on the datasets. The obtained K-PLS model will be helpful for the design of novel nicotine-like selective CYP2A6 inhibitors.

A Simulation-Based Comparison of Several Stochastic Linear Regression Methods in the Presence of Outliers.

ERIC Educational Resources Information Center

Rule, David L.

Several regression methods were examined within the framework of weighted structural regression (WSR), comparing their regression weight stability and score estimation accuracy in the presence of outlier contamination. The methods compared are: (1) ordinary least squares; (2) WSR ridge regression; (3) minimum risk regression; (4) minimum risk 2;…

Simultaneous determination of penicillin G salts by infrared spectroscopy: Evaluation of combining orthogonal signal correction with radial basis function-partial least squares regression

NASA Astrophysics Data System (ADS)

Talebpour, Zahra; Tavallaie, Roya; Ahmadi, Seyyed Hamid; Abdollahpour, Assem

2010-09-01

In this study, a new method for the simultaneous determination of penicillin G salts in pharmaceutical mixture via FT-IR spectroscopy combined with chemometrics was investigated. The mixture of penicillin G salts is a complex system due to similar analytical characteristics of components. Partial least squares (PLS) and radial basis function-partial least squares (RBF-PLS) were used to develop the linear and nonlinear relation between spectra and components, respectively. The orthogonal signal correction (OSC) preprocessing method was used to correct unexpected information, such as spectral overlapping and scattering effects. In order to compare the influence of OSC on PLS and RBF-PLS models, the optimal linear (PLS) and nonlinear (RBF-PLS) models based on conventional and OSC preprocessed spectra were established and compared. The obtained results demonstrated that OSC clearly enhanced the performance of both RBF-PLS and PLS calibration models. Also in the case of some nonlinear relation between spectra and component, OSC-RBF-PLS gave satisfactory results than OSC-PLS model which indicated that the OSC was helpful to remove extrinsic deviations from linearity without elimination of nonlinear information related to component. The chemometric models were tested on an external dataset and finally applied to the analysis commercialized injection product of penicillin G salts.

Paleotemperature reconstruction from mammalian phosphate δ18O records - an alternative view on data processing

NASA Astrophysics Data System (ADS)

Skrzypek, Grzegorz; Sadler, Rohan; Wiśniewski, Andrzej

2017-04-01

The stable oxygen isotope composition of phosphates (δ18O) extracted from mammalian bone and teeth material is commonly used as a proxy for paleotemperature. Historically, several different analytical and statistical procedures for determining air paleotemperatures from the measured δ18O of phosphates have been applied. This inconsistency in both stable isotope data processing and the application of statistical procedures has led to large and unwanted differences between calculated results. This study presents the uncertainty associated with two of the most commonly used regression methods: least squares inverted fit and transposed fit. We assessed the performance of these methods by designing and applying calculation experiments to multiple real-life data sets, calculating in reverse temperatures, and comparing them with true recorded values. Our calculations clearly show that the mean absolute errors are always substantially higher for the inverted fit (a causal model), with the transposed fit (a predictive model) returning mean values closer to the measured values (Skrzypek et al. 2015). The predictive models always performed better than causal models, with 12-65% lower mean absolute errors. Moreover, the least-squares regression (LSM) model is more appropriate than Reduced Major Axis (RMA) regression for calculating the environmental water stable oxygen isotope composition from phosphate signatures, as well as for calculating air temperature from the δ18O value of environmental water. The transposed fit introduces a lower overall error than the inverted fit for both the δ18O of environmental water and Tair calculations; therefore, the predictive models are more statistically efficient than the causal models in this instance. The direct comparison of paleotemperature results from different laboratories and studies may only be achieved if a single method of calculation is applied. Reference Skrzypek G., Sadler R., Wiśniewski A., 2016. Reassessment of recommendations for processing mammal phosphate δ18O data for paleotemperature reconstruction. Palaeogeography, Palaeoclimatology, Palaeoecology 446, 162-167.

Hyperspectral imaging for predicting the allicin and soluble solid content of garlic with variable selection algorithms and chemometric models.

PubMed

Rahman, Anisur; Faqeerzada, Mohammad A; Cho, Byoung-Kwan

2018-03-14

Allicin and soluble solid content (SSC) in garlic is the responsible for its pungent flavor and odor. However, current conventional methods such as the use of high-pressure liquid chromatography and a refractometer have critical drawbacks in that they are time-consuming, labor-intensive and destructive procedures. The present study aimed to predict allicin and SSC in garlic using hyperspectral imaging in combination with variable selection algorithms and calibration models. Hyperspectral images of 100 garlic cloves were acquired that covered two spectral ranges, from which the mean spectra of each clove were extracted. The calibration models included partial least squares (PLS) and least squares-support vector machine (LS-SVM) regression, as well as different spectral pre-processing techniques, from which the highest performing spectral preprocessing technique and spectral range were selected. Then, variable selection methods, such as regression coefficients, variable importance in projection (VIP) and the successive projections algorithm (SPA), were evaluated for the selection of effective wavelengths (EWs). Furthermore, PLS and LS-SVM regression methods were applied to quantitatively predict the quality attributes of garlic using the selected EWs. Of the established models, the SPA-LS-SVM model obtained an Rpred2 of 0.90 and standard error of prediction (SEP) of 1.01% for SSC prediction, whereas the VIP-LS-SVM model produced the best result with an Rpred2 of 0.83 and SEP of 0.19 mg g -1 for allicin prediction in the range 1000-1700 nm. Furthermore, chemical images of garlic were developed using the best predictive model to facilitate visualization of the spatial distributions of allicin and SSC. The present study clearly demonstrates that hyperspectral imaging combined with an appropriate chemometrics method can potentially be employed as a fast, non-invasive method to predict the allicin and SSC in garlic. © 2018 Society of Chemical Industry. © 2018 Society of Chemical Industry.

Modeling soil parameters using hyperspectral image reflectance in subtropical coastal wetlands

NASA Astrophysics Data System (ADS)

Anne, Naveen J. P.; Abd-Elrahman, Amr H.; Lewis, David B.; Hewitt, Nicole A.

2014-12-01

Developing spectral models of soil properties is an important frontier in remote sensing and soil science. Several studies have focused on modeling soil properties such as total pools of soil organic matter and carbon in bare soils. We extended this effort to model soil parameters in areas densely covered with coastal vegetation. Moreover, we investigated soil properties indicative of soil functions such as nutrient and organic matter turnover and storage. These properties include the partitioning of mineral and organic soil between particulate (>53 μm) and fine size classes, and the partitioning of soil carbon and nitrogen pools between stable and labile fractions. Soil samples were obtained from Avicennia germinans mangrove forest and Juncus roemerianus salt marsh plots on the west coast of central Florida. Spectra corresponding to field plot locations from Hyperion hyperspectral image were extracted and analyzed. The spectral information was regressed against the soil variables to determine the best single bands and optimal band combinations for the simple ratio (SR) and normalized difference index (NDI) indices. The regression analysis yielded levels of correlation for soil variables with R2 values ranging from 0.21 to 0.47 for best individual bands, 0.28 to 0.81 for two-band indices, and 0.53 to 0.96 for partial least-squares (PLS) regressions for the Hyperion image data. Spectral models using Hyperion data adequately (RPD > 1.4) predicted particulate organic matter (POM), silt + clay, labile carbon (C), and labile nitrogen (N) (where RPD = ratio of standard deviation to root mean square error of cross-validation [RMSECV]). The SR (0.53 μm, 2.11 μm) model of labile N with R2 = 0.81, RMSECV= 0.28, and RPD = 1.94 produced the best results in this study. Our results provide optimism that remote-sensing spectral models can successfully predict soil properties indicative of ecosystem nutrient and organic matter turnover and storage, and do so in areas with dense canopy cover.

Importance of spatial autocorrelation in modeling bird distributions at a continental scale

USGS Publications Warehouse

Bahn, V.; O'Connor, R.J.; Krohn, W.B.

2006-01-01

Spatial autocorrelation in species' distributions has been recognized as inflating the probability of a type I error in hypotheses tests, causing biases in variable selection, and violating the assumption of independence of error terms in models such as correlation or regression. However, it remains unclear whether these problems occur at all spatial resolutions and extents, and under which conditions spatially explicit modeling techniques are superior. Our goal was to determine whether spatial models were superior at large extents and across many different species. In addition, we investigated the importance of purely spatial effects in distribution patterns relative to the variation that could be explained through environmental conditions. We studied distribution patterns of 108 bird species in the conterminous United States using ten years of data from the Breeding Bird Survey. We compared the performance of spatially explicit regression models with non-spatial regression models using Akaike's information criterion. In addition, we partitioned the variance in species distributions into an environmental, a pure spatial and a shared component. The spatially-explicit conditional autoregressive regression models strongly outperformed the ordinary least squares regression models. In addition, partialling out the spatial component underlying the species' distributions showed that an average of 17% of the explained variation could be attributed to purely spatial effects independent of the spatial autocorrelation induced by the underlying environmental variables. We concluded that location in the range and neighborhood play an important role in the distribution of species. Spatially explicit models are expected to yield better predictions especially for mobile species such as birds, even in coarse-grained models with a large extent. ?? Ecography.

Radiomorphometric analysis of frontal sinus for sex determination.

PubMed

Verma, Saumya; Mahima, V G; Patil, Karthikeya

2014-09-01

Sex determination of unknown individuals carries crucial significance in forensic research, in cases where fragments of skull persist with no likelihood of identification based on dental arch. In these instances sex determination becomes important to rule out certain number of possibilities instantly and helps in establishing a biological profile of human remains. The aim of the study is to evaluate a mathematical method based on logistic regression analysis capable of ascertaining the sex of individuals in the South Indian population. The study was conducted in the department of Oral Medicine and Radiology. The right and left areas, maximum height, width of frontal sinus were determined in 100 Caldwell views of 50 women and 50 men aged 20 years and above, with the help of Vernier callipers and a square grid with 1 square measuring 1mm(2) in area. Student's t-test, logistic regression analysis. The mean values of variables were greater in men, based on Student's t-test at 5% level of significance. The mathematical model based on logistic regression analysis gave percentage agreement of total area to correctly predict the female gender as 55.2%, of right area as 60.9% and of left area as 55.2%. The areas of the frontal sinus and the logistic regression proved to be unreliable in sex determination. (Logit = 0.924 - 0.00217 × right area).

Delineation and Prediction Uncertainty of Areas Contributing Recharge to Selected Well Fields in Wetland and Coastal Settings, Southern Rhode Island

USGS Publications Warehouse

Friesz, Paul J.

2010-01-01

Areas contributing recharge to four well fields in two study sites in southern Rhode Island were delineated on the basis of steady-state groundwater-flow models representing average hydrologic conditions. The wells are screened in sand and gravel deposits in wetland and coastal settings. The groundwater-flow models were calibrated by inverse modeling using nonlinear regression. Summary statistics from nonlinear regression were used to evaluate the uncertainty associated with the predicted areas contributing recharge to the well fields. In South Kingstown, two United Water Rhode Island well fields are in Mink Brook watershed and near Worden Pond and extensive wetlands. Wetland deposits of peat near the well fields generally range in thickness from 5 to 8 feet. Analysis of water-level drawdowns in a piezometer screened beneath the peat during a 20-day pumping period indicated vertical leakage and a vertical hydraulic conductivity for the peat of roughly 0.01 ft/d. The simulated area contributing recharge for average withdrawals of 2,138 gallons per minute during 2003-07 extended to groundwater divides in mostly till and morainal deposits, and it encompassed 2.30 square miles. Most of a sand and gravel mining operation between the well fields was in the simulated contributing area. For the maximum pumping capacity (5,100 gallons per minute), the simulated area contributing recharge expanded to 5.54 square miles. The well fields intercepted most of the precipitation recharge in Mink Brook watershed and in an adjacent small watershed, and simulated streams ceased to flow. The simulated contributing area to the well fields included an area beneath Worden Pond and a remote, isolated area in upland till on the opposite side of Worden Pond from the well fields. About 12 percent of the pumped water was derived from Worden Pond. In Charlestown, the Central Beach Fire District and the East Beach Water Association well fields are on a small (0.85 square mile) peninsula in a coastal setting. The wells are screened in a coarse-grained, ice-proximal part of a morphosequence with saturated thicknesses generally less than 30 feet on the peninsula. The simulated area contributing recharge for the average withdrawal (16 gallons per minute) during 2003-07 was 0.018 square mile. The contributing area extended southwestward from the well fields to a simulated groundwater mound; it underlay part of a small nearby wetland, and it included isolated areas on the side of the wetland opposite the well fields. For the maximum pumping rate (230 gallons per minute), the simulated area contributing recharge (0.26 square mile) expanded in all directions; it included a till area on the peninsula, and it underlay part of a nearby pond. Because the well fields are screened in a thin aquifer, simulated groundwater traveltimes from recharge locations to the discharging wells were short: 94 percent of the traveltimes were 10 years or less, and the median traveltime was 1.3 years. Model-prediction uncertainty was evaluated using a Monte Carlo analysis; the parameter variance-covariance matrix from nonlinear regression was used to create parameter sets for the analysis. Important parameters for model prediction that could not be estimated by nonlinear regression were incorporated into the variance-covariance matrix. For the South Kingstown study site, observations provided enough information to constrain the uncertainty of these parameters within realistic ranges, but for the Charlestown study site, prior information on parameters was required. Thus, the uncertainty analysis for the South Kingstown study site was an outcome of calibrating the model to available observations, but the Charlestown study site was also dependent on information provided by the modeler. A water budget and model-fit statistical criteria were used to assess parameter sets so that prediction uncertainty was not overestimated. For the scenarios using maximum pumping rates at both study

Identification and Severity Determination of Wheat Stripe Rust and Wheat Leaf Rust Based on Hyperspectral Data Acquired Using a Black-Paper-Based Measuring Method.

PubMed

Wang, Hui; Qin, Feng; Ruan, Liu; Wang, Rui; Liu, Qi; Ma, Zhanhong; Li, Xiaolong; Cheng, Pei; Wang, Haiguang

2016-01-01

It is important to implement detection and assessment of plant diseases based on remotely sensed data for disease monitoring and control. Hyperspectral data of healthy leaves, leaves in incubation period and leaves in diseased period of wheat stripe rust and wheat leaf rust were collected under in-field conditions using a black-paper-based measuring method developed in this study. After data preprocessing, the models to identify the diseases were built using distinguished partial least squares (DPLS) and support vector machine (SVM), and the disease severity inversion models of stripe rust and the disease severity inversion models of leaf rust were built using quantitative partial least squares (QPLS) and support vector regression (SVR). All the models were validated by using leave-one-out cross validation and external validation. The diseases could be discriminated using both distinguished partial least squares and support vector machine with the accuracies of more than 99%. For each wheat rust, disease severity levels were accurately retrieved using both the optimal QPLS models and the optimal SVR models with the coefficients of determination (R2) of more than 0.90 and the root mean square errors (RMSE) of less than 0.15. The results demonstrated that identification and severity evaluation of stripe rust and leaf rust at the leaf level could be implemented based on the hyperspectral data acquired using the developed method. A scientific basis was provided for implementing disease monitoring by using aerial and space remote sensing technologies.

Identification and Severity Determination of Wheat Stripe Rust and Wheat Leaf Rust Based on Hyperspectral Data Acquired Using a Black-Paper-Based Measuring Method

PubMed Central

Ruan, Liu; Wang, Rui; Liu, Qi; Ma, Zhanhong; Li, Xiaolong; Cheng, Pei; Wang, Haiguang

2016-01-01

It is important to implement detection and assessment of plant diseases based on remotely sensed data for disease monitoring and control. Hyperspectral data of healthy leaves, leaves in incubation period and leaves in diseased period of wheat stripe rust and wheat leaf rust were collected under in-field conditions using a black-paper-based measuring method developed in this study. After data preprocessing, the models to identify the diseases were built using distinguished partial least squares (DPLS) and support vector machine (SVM), and the disease severity inversion models of stripe rust and the disease severity inversion models of leaf rust were built using quantitative partial least squares (QPLS) and support vector regression (SVR). All the models were validated by using leave-one-out cross validation and external validation. The diseases could be discriminated using both distinguished partial least squares and support vector machine with the accuracies of more than 99%. For each wheat rust, disease severity levels were accurately retrieved using both the optimal QPLS models and the optimal SVR models with the coefficients of determination (R2) of more than 0.90 and the root mean square errors (RMSE) of less than 0.15. The results demonstrated that identification and severity evaluation of stripe rust and leaf rust at the leaf level could be implemented based on the hyperspectral data acquired using the developed method. A scientific basis was provided for implementing disease monitoring by using aerial and space remote sensing technologies. PMID:27128464

Development of variable pathlength UV-vis spectroscopy combined with partial-least-squares regression for wastewater chemical oxygen demand (COD) monitoring.

PubMed

Chen, Baisheng; Wu, Huanan; Li, Sam Fong Yau

2014-03-01

To overcome the challenging task to select an appropriate pathlength for wastewater chemical oxygen demand (COD) monitoring with high accuracy by UV-vis spectroscopy in wastewater treatment process, a variable pathlength approach combined with partial-least squares regression (PLSR) was developed in this study. Two new strategies were proposed to extract relevant information of UV-vis spectral data from variable pathlength measurements. The first strategy was by data fusion with two data fusion levels: low-level data fusion (LLDF) and mid-level data fusion (MLDF). Predictive accuracy was found to improve, indicated by the lower root-mean-square errors of prediction (RMSEP) compared with those obtained for single pathlength measurements. Both fusion levels were found to deliver very robust PLSR models with residual predictive deviations (RPD) greater than 3 (i.e. 3.22 and 3.29, respectively). The second strategy involved calculating the slopes of absorbance against pathlength at each wavelength to generate slope-derived spectra. Without the requirement to select the optimal pathlength, the predictive accuracy (RMSEP) was improved by 20-43% as compared to single pathlength spectroscopy. Comparing to nine-factor models from fusion strategy, the PLSR model from slope-derived spectroscopy was found to be more parsimonious with only five factors and more robust with residual predictive deviation (RPD) of 3.72. It also offered excellent correlation of predicted and measured COD values with R(2) of 0.936. In sum, variable pathlength spectroscopy with the two proposed data analysis strategies proved to be successful in enhancing prediction performance of COD in wastewater and showed high potential to be applied in on-line water quality monitoring. Copyright © 2013 Elsevier B.V. All rights reserved.

Detecting outliers when fitting data with nonlinear regression – a new method based on robust nonlinear regression and the false discovery rate

PubMed Central

Motulsky, Harvey J; Brown, Ronald E

2006-01-01

Background Nonlinear regression, like linear regression, assumes that the scatter of data around the ideal curve follows a Gaussian or normal distribution. This assumption leads to the familiar goal of regression: to minimize the sum of the squares of the vertical or Y-value distances between the points and the curve. Outliers can dominate the sum-of-the-squares calculation, and lead to misleading results. However, we know of no practical method for routinely identifying outliers when fitting curves with nonlinear regression. Results We describe a new method for identifying outliers when fitting data with nonlinear regression. We first fit the data using a robust form of nonlinear regression, based on the assumption that scatter follows a Lorentzian distribution. We devised a new adaptive method that gradually becomes more robust as the method proceeds. To define outliers, we adapted the false discovery rate approach to handling multiple comparisons. We then remove the outliers, and analyze the data using ordinary least-squares regression. Because the method combines robust regression and outlier removal, we call it the ROUT method. When analyzing simulated data, where all scatter is Gaussian, our method detects (falsely) one or more outlier in only about 1–3% of experiments. When analyzing data contaminated with one or several outliers, the ROUT method performs well at outlier identification, with an average False Discovery Rate less than 1%. Conclusion Our method, which combines a new method of robust nonlinear regression with a new method of outlier identification, identifies outliers from nonlinear curve fits with reasonable power and few false positives. PMID:16526949

Methods for estimating annual exceedance-probability streamflows for streams in Kansas based on data through water year 2015

USGS Publications Warehouse

Painter, Colin C.; Heimann, David C.; Lanning-Rush, Jennifer L.

2017-08-14

A study was done by the U.S. Geological Survey in cooperation with the Kansas Department of Transportation and the Federal Emergency Management Agency to develop regression models to estimate peak streamflows of annual exceedance probabilities of 50, 20, 10, 4, 2, 1, 0.5, and 0.2 percent at ungaged locations in Kansas. Peak streamflow frequency statistics from selected streamgages were related to contributing drainage area and average precipitation using generalized least-squares regression analysis. The peak streamflow statistics were derived from 151 streamgages with at least 25 years of streamflow data through 2015. The developed equations can be used to predict peak streamflow magnitude and frequency within two hydrologic regions that were defined based on the effects of irrigation. The equations developed in this report are applicable to streams in Kansas that are not substantially affected by regulation, surface-water diversions, or urbanization. The equations are intended for use for streams with contributing drainage areas ranging from 0.17 to 14,901 square miles in the nonirrigation effects region and, 1.02 to 3,555 square miles in the irrigation-affected region, corresponding to the range of drainage areas of the streamgages used in the development of the regional equations.

Harvest-time prediction of apple physiological indices using fiber optic Fourier transform near-infrared spectrometer

NASA Astrophysics Data System (ADS)

Liu, Yande; Ying, Yibin; Lu, Huishan; Fu, Xiaping

2004-12-01

This work evaluates the feasibility of Fourier transform near infrared (FT-NIR) spectrometry for rapid determining the total soluble solids content and acidity of apple fruit. Intact apple fruit were measured by reflectance FT-NIR in 800-2500 nm range. FT-NIR models were developed based on partial least square (PLS) regression and principal component regress (PCR) with respect to the reflectance and its first derivative, the logarithms of the reflectance reciprocal and its second derivative. The above regression models, related the FT-NIR spectra to soluble solids content (SSC), titratable acidity (TA) and available acidity (pH). The best combination, based on the prediction results, was PLS models with respect to the logarithms of the reflectance reciprocal. Predictions with PLS models resulted standard errors of prediction (SEP) of 0.455, 0.044 and 0.068, and correlation coefficients of 0.968, 0.728 and 0.831 for SSC, TA and pH, respectively. It was concluded that by using the FT-NIR spectrometry measurement system, in the appropriate spectral range, it is possible to nondestructively assess the maturity factors of apple fruit.

Correlations of turbidity to suspended-sediment concentration in the Toutle River Basin, near Mount St. Helens, Washington, 2010-11

USGS Publications Warehouse

Uhrich, Mark A.; Kolasinac, Jasna; Booth, Pamela L.; Fountain, Robert L.; Spicer, Kurt R.; Mosbrucker, Adam R.

2014-01-01

Researchers at the U.S. Geological Survey, Cascades Volcano Observatory, investigated alternative methods for the traditional sample-based sediment record procedure in determining suspended-sediment concentration (SSC) and discharge. One such sediment-surrogate technique was developed using turbidity and discharge to estimate SSC for two gaging stations in the Toutle River Basin near Mount St. Helens, Washington. To provide context for the study, methods for collecting sediment data and monitoring turbidity are discussed. Statistical methods used include the development of ordinary least squares regression models for each gaging station. Issues of time-related autocorrelation also are evaluated. Addition of lagged explanatory variables was used to account for autocorrelation in the turbidity, discharge, and SSC data. Final regression model equations and plots are presented for the two gaging stations. The regression models support near-real-time estimates of SSC and improved suspended-sediment discharge records by incorporating continuous instream turbidity. Future use of such models may potentially lower the costs of sediment monitoring by reducing time it takes to collect and process samples and to derive a sediment-discharge record.

Measurement of process variables in solid-state fermentation of wheat straw using FT-NIR spectroscopy and synergy interval PLS algorithm.

PubMed

Jiang, Hui; Liu, Guohai; Mei, Congli; Yu, Shuang; Xiao, Xiahong; Ding, Yuhan

2012-11-01

The feasibility of rapid determination of the process variables (i.e. pH and moisture content) in solid-state fermentation (SSF) of wheat straw using Fourier transform near infrared (FT-NIR) spectroscopy was studied. Synergy interval partial least squares (siPLS) algorithm was implemented to calibrate regression model. The number of PLS factors and the number of subintervals were optimized simultaneously by cross-validation. The performance of the prediction model was evaluated according to the root mean square error of cross-validation (RMSECV), the root mean square error of prediction (RMSEP) and the correlation coefficient (R). The measurement results of the optimal model were obtained as follows: RMSECV=0.0776, R(c)=0.9777, RMSEP=0.0963, and R(p)=0.9686 for pH model; RMSECV=1.3544% w/w, R(c)=0.8871, RMSEP=1.4946% w/w, and R(p)=0.8684 for moisture content model. Finally, compared with classic PLS and iPLS models, the siPLS model revealed its superior performance. The overall results demonstrate that FT-NIR spectroscopy combined with siPLS algorithm can be used to measure process variables in solid-state fermentation of wheat straw, and NIR spectroscopy technique has a potential to be utilized in SSF industry. Copyright © 2012 Elsevier B.V. All rights reserved.

Modeling of surface dust concentrations using neural networks and kriging

NASA Astrophysics Data System (ADS)

Buevich, Alexander G.; Medvedev, Alexander N.; Sergeev, Alexander P.; Tarasov, Dmitry A.; Shichkin, Andrey V.; Sergeeva, Marina V.; Atanasova, T. B.

2016-12-01

Creating models which are able to accurately predict the distribution of pollutants based on a limited set of input data is an important task in environmental studies. In the paper two neural approaches: (multilayer perceptron (MLP)) and generalized regression neural network (GRNN)), and two geostatistical approaches: (kriging and cokriging), are using for modeling and forecasting of dust concentrations in snow cover. The area of study is under the influence of dust emissions from a copper quarry and a several industrial companies. The comparison of two mentioned approaches is conducted. Three indices are used as the indicators of the models accuracy: the mean absolute error (MAE), root mean square error (RMSE) and relative root mean square error (RRMSE). Models based on artificial neural networks (ANN) have shown better accuracy. When considering all indices, the most precision model was the GRNN, which uses as input parameters for modeling the coordinates of sampling points and the distance to the probable emissions source. The results of work confirm that trained ANN may be more suitable tool for modeling of dust concentrations in snow cover.

A fast and direct spectrophotometric method for the simultaneous determination of methyl paraben and hydroquinone in cosmetic products using successive projections algorithm.

PubMed

Esteki, M; Nouroozi, S; Shahsavari, Z

2016-02-01

To develop a simple and efficient spectrophotometric technique combined with chemometrics for the simultaneous determination of methyl paraben (MP) and hydroquinone (HQ) in cosmetic products, and specifically, to: (i) evaluate the potential use of successive projections algorithm (SPA) to derivative spectrophotometric data in order to provide sufficient accuracy and model robustness and (ii) determine MP and HQ concentration in cosmetics without tedious pre-treatments such as derivatization or extraction techniques which are time-consuming and require hazardous solvents. The absorption spectra were measured in the wavelength range of 200-350 nm. Prior to performing chemometric models, the original and first-derivative absorption spectra of binary mixtures were used as calibration matrices. Variable selected by successive projections algorithm was used to obtain multiple linear regression (MLR) models based on a small subset of wavelengths. The number of wavelengths and the starting vector were optimized, and the comparison of the root mean square error of calibration (RMSEC) and cross-validation (RMSECV) was applied to select effective wavelengths with the least collinearity and redundancy. Principal component regression (PCR) and partial least squares (PLS) were also developed for comparison. The concentrations of the calibration matrix ranged from 0.1 to 20 μg mL(-1) for MP, and from 0.1 to 25 μg mL(-1) for HQ. The constructed models were tested on an external validation data set and finally cosmetic samples. The results indicated that successive projections algorithm-multiple linear regression (SPA-MLR), applied on the first-derivative spectra, achieved the optimal performance for two compounds when compared with the full-spectrum PCR and PLS. The root mean square error of prediction (RMSEP) was 0.083, 0.314 for MP and HQ, respectively. To verify the accuracy of the proposed method, a recovery study on real cosmetic samples was carried out with satisfactory results (84-112%). The proposed method, which is an environmentally friendly approach, using minimum amount of solvent, is a simple, fast and low-cost analysis method that can provide high accuracy and robust models. The suggested method does not need any complex extraction procedure which is time-consuming and requires hazardous solvents. © 2015 Society of Cosmetic Scientists and the Société Française de Cosmétologie.

Comparison of random regression models with Legendre polynomials and linear splines for production traits and somatic cell score of Canadian Holstein cows.

PubMed

Bohmanova, J; Miglior, F; Jamrozik, J; Misztal, I; Sullivan, P G

2008-09-01

A random regression model with both random and fixed regressions fitted by Legendre polynomials of order 4 was compared with 3 alternative models fitting linear splines with 4, 5, or 6 knots. The effects common for all models were a herd-test-date effect, fixed regressions on days in milk (DIM) nested within region-age-season of calving class, and random regressions for additive genetic and permanent environmental effects. Data were test-day milk, fat and protein yields, and SCS recorded from 5 to 365 DIM during the first 3 lactations of Canadian Holstein cows. A random sample of 50 herds consisting of 96,756 test-day records was generated to estimate variance components within a Bayesian framework via Gibbs sampling. Two sets of genetic evaluations were subsequently carried out to investigate performance of the 4 models. Models were compared by graphical inspection of variance functions, goodness of fit, error of prediction of breeding values, and stability of estimated breeding values. Models with splines gave lower estimates of variances at extremes of lactations than the model with Legendre polynomials. Differences among models in goodness of fit measured by percentages of squared bias, correlations between predicted and observed records, and residual variances were small. The deviance information criterion favored the spline model with 6 knots. Smaller error of prediction and higher stability of estimated breeding values were achieved by using spline models with 5 and 6 knots compared with the model with Legendre polynomials. In general, the spline model with 6 knots had the best overall performance based upon the considered model comparison criteria.

Limited Sampling Strategy for Accurate Prediction of Pharmacokinetics of Saroglitazar: A 3-point Linear Regression Model Development and Successful Prediction of Human Exposure.

PubMed

Joshi, Shuchi N; Srinivas, Nuggehally R; Parmar, Deven V

2018-03-01

Our aim was to develop and validate the extrapolative performance of a regression model using a limited sampling strategy for accurate estimation of the area under the plasma concentration versus time curve for saroglitazar. Healthy subject pharmacokinetic data from a well-powered food-effect study (fasted vs fed treatments; n = 50) was used in this work. The first 25 subjects' serial plasma concentration data up to 72 hours and corresponding AUC 0-t (ie, 72 hours) from the fasting group comprised a training dataset to develop the limited sampling model. The internal datasets for prediction included the remaining 25 subjects from the fasting group and all 50 subjects from the fed condition of the same study. The external datasets included pharmacokinetic data for saroglitazar from previous single-dose clinical studies. Limited sampling models were composed of 1-, 2-, and 3-concentration-time points' correlation with AUC 0-t of saroglitazar. Only models with regression coefficients (R 2 ) >0.90 were screened for further evaluation. The best R 2 model was validated for its utility based on mean prediction error, mean absolute prediction error, and root mean square error. Both correlations between predicted and observed AUC 0-t of saroglitazar and verification of precision and bias using Bland-Altman plot were carried out. None of the evaluated 1- and 2-concentration-time points models achieved R 2 > 0.90. Among the various 3-concentration-time points models, only 4 equations passed the predefined criterion of R 2 > 0.90. Limited sampling models with time points 0.5, 2, and 8 hours (R 2 = 0.9323) and 0.75, 2, and 8 hours (R 2 = 0.9375) were validated. Mean prediction error, mean absolute prediction error, and root mean square error were <30% (predefined criterion) and correlation (r) was at least 0.7950 for the consolidated internal and external datasets of 102 healthy subjects for the AUC 0-t prediction of saroglitazar. The same models, when applied to the AUC 0-t prediction of saroglitazar sulfoxide, showed mean prediction error, mean absolute prediction error, and root mean square error <30% and correlation (r) was at least 0.9339 in the same pool of healthy subjects. A 3-concentration-time points limited sampling model predicts the exposure of saroglitazar (ie, AUC 0-t ) within predefined acceptable bias and imprecision limit. Same model was also used to predict AUC 0-∞ . The same limited sampling model was found to predict the exposure of saroglitazar sulfoxide within predefined criteria. This model can find utility during late-phase clinical development of saroglitazar in the patient population. Copyright © 2018 Elsevier HS Journals, Inc. All rights reserved.

NARMAX model identification of a palm oil biodiesel engine using multi-objective optimization differential evolution

NASA Astrophysics Data System (ADS)

Mansor, Zakwan; Zakaria, Mohd Zakimi; Nor, Azuwir Mohd; Saad, Mohd Sazli; Ahmad, Robiah; Jamaluddin, Hishamuddin

2017-09-01

This paper presents the black-box modelling of palm oil biodiesel engine (POB) using multi-objective optimization differential evolution (MOODE) algorithm. Two objective functions are considered in the algorithm for optimization; minimizing the number of term of a model structure and minimizing the mean square error between actual and predicted outputs. The mathematical model used in this study to represent the POB system is nonlinear auto-regressive moving average with exogenous input (NARMAX) model. Finally, model validity tests are applied in order to validate the possible models that was obtained from MOODE algorithm and lead to select an optimal model.

Meteorological adjustment of yearly mean values for air pollutant concentration comparison

NASA Technical Reports Server (NTRS)

Sidik, S. M.; Neustadter, H. E.

1976-01-01

Using multiple linear regression analysis, models which estimate mean concentrations of Total Suspended Particulate (TSP), sulfur dioxide, and nitrogen dioxide as a function of several meteorologic variables, two rough economic indicators, and a simple trend in time are studied. Meteorologic data were obtained and do not include inversion heights. The goodness of fit of the estimated models is partially reflected by the squared coefficient of multiple correlation which indicates that, at the various sampling stations, the models accounted for about 23 to 47 percent of the total variance of the observed TSP concentrations. If the resulting model equations are used in place of simple overall means of the observed concentrations, there is about a 20 percent improvement in either: (1) predicting mean concentrations for specified meteorological conditions; or (2) adjusting successive yearly averages to allow for comparisons devoid of meteorological effects. An application to source identification is presented using regression coefficients of wind velocity predictor variables.

ELASTIC NET FOR COX'S PROPORTIONAL HAZARDS MODEL WITH A SOLUTION PATH ALGORITHM.

PubMed

Wu, Yichao

2012-01-01

For least squares regression, Efron et al. (2004) proposed an efficient solution path algorithm, the least angle regression (LAR). They showed that a slight modification of the LAR leads to the whole LASSO solution path. Both the LAR and LASSO solution paths are piecewise linear. Recently Wu (2011) extended the LAR to generalized linear models and the quasi-likelihood method. In this work we extend the LAR further to handle Cox's proportional hazards model. The goal is to develop a solution path algorithm for the elastic net penalty (Zou and Hastie (2005)) in Cox's proportional hazards model. This goal is achieved in two steps. First we extend the LAR to optimizing the log partial likelihood plus a fixed small ridge term. Then we define a path modification, which leads to the solution path of the elastic net regularized log partial likelihood. Our solution path is exact and piecewise determined by ordinary differential equation systems.

Physician leadership styles and effectiveness: an empirical study.

PubMed

Xirasagar, Sudha; Samuels, Michael E; Stoskopf, Carleen H

2005-12-01

The authors study the association between physician leadership styles and leadership effectiveness. Executive directors of community health centers were surveyed (269 respondents; response rate = 40.9 percent) for their perceptions of the medical director's leadership behaviors and effectiveness, using an adapted Multifactor Leadership Questionnaire (43 items on a 0-4 point Likert-type scale), with additional questions on demographics and the center's clinical goals and achievements. The authors hypothesize that transformational leadership would be more positively associated with executive directors' ratings of effectiveness, satisfaction with the leader, and subordinate extra effort, as well as the center's clinical goal achievement, than transactional or laissez-faire leadership. Separate ordinary least squares regressions were used to model each of the effectiveness measures, and general linear model regression was used to model clinical goal achievement. Results support the hypothesis and suggest that physician leadership development using the transformational leadership model may result in improved health care quality and cost control.

Outcome modelling strategies in epidemiology: traditional methods and basic alternatives

PubMed Central

Greenland, Sander; Daniel, Rhian; Pearce, Neil

2016-01-01

Abstract Controlling for too many potential confounders can lead to or aggravate problems of data sparsity or multicollinearity, particularly when the number of covariates is large in relation to the study size. As a result, methods to reduce the number of modelled covariates are often deployed. We review several traditional modelling strategies, including stepwise regression and the ‘change-in-estimate’ (CIE) approach to deciding which potential confounders to include in an outcome-regression model for estimating effects of a targeted exposure. We discuss their shortcomings, and then provide some basic alternatives and refinements that do not require special macros or programming. Throughout, we assume the main goal is to derive the most accurate effect estimates obtainable from the data and commercial software. Allowing that most users must stay within standard software packages, this goal can be roughly approximated using basic methods to assess, and thereby minimize, mean squared error (MSE). PMID:27097747