A general procedure for evaluating models and ensemble Support Vector Regression

Calibrating lab and field reflectance spectra for nutrient estimation in potato plants using local support vector regression models

The present study quantitative structure-property relationship (QSPR) model developed for the adsorption efficiency (AE) of 70 pesticides in water sample on MWCNT/PbO2 solid phase extraction cartridge. Stepwise-multiple linear regression (SW-MLR) method employed for selection of descriptors. The selected descriptors are MATS7v, MATS6c, GATS3s, ATSC6i, C040, SpMin8_Bhi, E2v, JGI1 and Mor08u. Curiosity at the effective descriptors indicates electronic, topological and geometrical characteristics of studied pesticides are the most effective parameters on their AE on MWCNT/PbO2 nanocomposite adsorbent. Symbolic regression via genetic programming (SR-GP) utilized to offer the symbolic regression QSPR model. The accuracy and predictive power of the SR-GP model compared with traditional linear and nonlinear regression models contain multiple linear regression (MLR) and support vector regression (SVR). Inspection the fitness parameters confirmed the superiority of SR-GP model over MLR, and SVR models. In SR-GP model, the correlation coefficient (R) was 0.930 and 0.890, and the root mean square errors (RMSE) were 0.04 and 0.05 for the training and test sets, respectively. These results can be used to predict the AE for other pesticides by MWCNT/PbO2 adsorbent and designing a more efficient nano cartridge for SPE.

Permeability prediction of isolated channel sands using machine learning

Feed efficiency traits are important for reducing feeding costs while potentially decreasing the environmental impact of broiler production. This study aimed to compare the predictive ability of different genome-wide prediction models for residual feed intake (RFI) calculated with linear and non-linear regression techniques. We also evaluated the heritability of such RFI as well as their associations with body weight. Genotypes and phenotypes used in this study were recorded from 32,747 purebred broiler chickens of both sexes. The individual feed intake was monitored daily in floor-raised birds using a radio-frequency antenna system. RFI was computed as the residual term of total feed intake (FI) adjusted for the average metabolic weight (AMW), body weight gain (BWG), and sex. The following regression methods were used to compute RFI: multiple linear regression (MLR), generalized additive model using splines (GAM), support vector regression using a radial basis kernel (SVR), and two-hidden layer neural network (NN). The different RFI assessments were fitted jointly with body weight measured at the end of the feeding trial (FBW) using a two-trait mixed model. The effects of contemporary groups were fitted as fixed for both traits, whereas sex was also considered for FBW. RFI (computed with different regression techniques) was predicted using different genomic prediction models (BayesA, BayesB, BayesCπ, GBLUP, RKHS, and SVR). Genomic prediction accuracy (ACC) was assessed using a random forward validation scheme, in which data from previous generations were used for training the models while later generations were used to test predictions. The effect of the RFI computation method on ACC was tested using an unpaired Wilcoxon test. Sample variance observed for RFI ranged between 57,207.04 and 59,254.76, depending on the regression method used to compute it. Overall, the RFI sample variance was smaller for non-linear methods (NN, SVR, and GAM) compared with the linear model (MLR). The regression methods did not impact the genetic parameter estimates for RFI, which presented a heritability value around 0.32 ± 0.015 and genetic correlation with FBW ranging between 0.22 ± 0.04 and 0.25 ± 0.04. Prediction accuracy across genomic regression methods was slightly higher (P-value = 0.038) when RFI was computed with SVR (0.336 ± 0.023) compared with the MLR (0.329 ± 0.025), although this small difference would have little practical implications. No differences in terms of ACC were observed among the other regression methods used to compute RFI. Our preliminary results indicate that the regression technique used to compute RFI does not have major implications for this trait. Therefore, the MLR method is the preferred choice as it is simpler to implement, and it provides easy-to-interpret parameters.

Processing experimental data with uncertainty is a huge challenge for robust regression modeling and high-accuracy data forecasting. The objective of the present study is the development of a robust support vector regression model for processing observed data of independent variables containing polyhedral perturbation and achievement of high forecasting precision. Firstly, the conception of the data collection of polyhedral perturbation is given and analyzed. Secondly, a novel robust support vector regression model is constructed by replacing the original independent variables in a linear support vector regression model with the independent variables containing polyhedral perturbation. Thirdly, the novel robust linear support vector regression method is also expanded to the non-linear regression model. Both the two models are validated by the linear and nonlinear numerical regression experiments. Comparison of the experimental results show that the proposed method provide more accurate predictions than the traditional regression method for processing data with polyhedral perturbation.

Exploring different quantum chemical quantities for lead compounds is an ongoing approach in identifying crucial structural activity related features that are contributing into their biological activities. Herein, activity-related quantum chemical calculations were performed for the selected estrogenic stilbene derivatives using density functional theory (DFT) with B3LYP functional and 6-311++G** basis set. In addition, specific activity-related geometry-independent drug-like properties are discussed for these derivatives. To obtain the mathematical model that correlates the chemical descriptors with their measured estrogenic activities, the quantitative structure activity relationship (QSAR) is established using multiple linear regression (MLR) and support vector regression (SVR) methods. Satisfactory fit with a reasonable regression correlation coefficient (${\rm{R}}^{2}= 0.78$R2=0.78) between predicted and experimental $pEC_{50}$pEC50 values is observed using MLR method. The present study identifies the essential physicochemical descriptors that effectively contribute in the estrogenic activity. The applied approach provides helpful insight into the designing novel estrogenic agents with improved anticancer activities.

Linear regression (LR) and support vector regression (SVR) are widely used in data analysis. Geometrical correlation learning (GcLearn) was proposed recently to improve the predictive ability of LR and SVR through mining and using correlations between data of a variable (inner correlation). This paper theoretically analyzes prediction performance of the GcLearn method and proves that GcLearn LR and SVR will have better prediction performance than traditional LR and SVR for prediction tasks when good inner correlations are obtained and predictions by traditional LR and SVR are far away from their neighbor training data under inner correlation. This gives the applicable condition of GcLearn method.

Crops monitoring is a challengeable subject that radar images can help it. The applicability of Sentinel-1 SAR data with dual polarization provided a splendid opportunity to develop a method for estimating rice parameters. Heights of cereal and biomass are two significant characteristics of rice that can be estimated with assessing satellite data and field measurements by classical regression methods [multiple linear regression (MLR), relevance vector regression (RVR), and support vector regression (SVR)]. In this study, Sentinel-1 SAR data from April 2018 to September 2018 in Astaneh-ye Ashrafiyeh region in the north of Iran were used. To evaluate and analyze validation of regression methods, field measurements (gathered from 15 plots) were utilized. The efficiency of nonparametric methods (SVR and RVR) is much better than that of the parametric regression (MLR) for rice parameter estimations. Among nonparametric approaches, RVR method has better results than SVR, because of the highest correlation coefficient (R2) and lowest root mean square error (RMSE). R2 = 0.92, RMSE = 162.1, and MAE = 971.9 and R2 = 0.92, RMSE = 10.9, and MAE = 70.71 are the results of height and biomass, respectively.

In order to test the applicability of support vector regression (SVR) in near-infrared (NIR) spectral analysis of tobacco,NIR calibration model for starch in tobacco were developed by processing NIR spectra and chemically determined starch content data of 187 tobacco samples with SVR,partial least square regression (PLS),multiplicative linear regression (MLR) and error back propagation artificial neural network (BP-ANN). The obtained model was tested through internal validation and external validation with leave-one-out cross validation (LOOCV) and independent sample set. The results showed that the accuracy of SVR model was slightly higher than that of BP-ANN,PLS and MLR models,it implied that SVR algorithm could be applied to NIR analysis of starch in tobacco.

There are numerous approaches for predicting faults in the software engineering research field. Software entropy metrics introduced by Hassan (Predicting faults using the complexity of code changes, 78–88, 2009) [1] are also popularly used for fault prediction. In previous studies, statistical linear regression (SLR) and support vector regression (SVR) for predicting faults using software entropy metrics have been validated. However, other machine learning approaches have not yet been explored. This study explores the applicability of locally weighted regression (LWR) approach for predicting faults using the software entropy metrics and compares it with SVR. It is noticed that the LWR performs better than SVR in most of the cases.

Asphaltene aggregation affects the entire production chain of the petrochemical industry, which also poses environmental challenges for oil pollution remediation. The aggregation process has been investigated for decades, but it remains unclear how the free energy of asphaltene association in solvents is correlated to its molecular structure. In this study, dimerization energies of 28 types of asphaltenes in water and toluene were calculated using the umbrella sampling method. Structural parameters related to the atom types and functional groups were screened to identify the factors most influencing the dimerization energy using multiple linear regression, multilayer perceptron, and support vector regression. Results demonstrated that the influence of molecular structure on asphaltene association in water was nonlinear, while attempts to capture the relationship using linear regression had larger error. The linkage per aromatic ring, number of aromatic carbons, and number of aliphatic chains were the top ...

The emission of blast furnace (BF) exhaust gas has been criticized by society. It is momentous to quickly predict the comprehensive coke ratio (CCR) of BF, because CCR is one of the important indicators for evaluating gas emissions, energy consumption, and production stability, and also affects composite economic benefits. In this article, 13 data-driven prediction techniques, including six conventional and seven ensemble methods, are applied to predict CCR. The result of ten-fold cross-validation indicates that multiple linear regression (MLR) and support vector regression (SVR) based on radial basis function are superior to the other methods. The mean absolute error, the root mean square error, and the coefficient of determination (R 2) of the MLR model are 1.079 kg·t−1, 1.668, and 0.973, respectively. The three indicators of the SVR model are 1.158 kg·t−1, 1.878, and 0.975, respectively. Furthermore, AdaBoost based on linear regression has also strong prediction ability and generalization performance. The three methods have important significances both in theory and in practice for predicting CCR. Moreover, the models constructed here can provide valuable hints into realizing data-driven control of the BF process.

With the vast availability of traffic sensing data on highway, real-time traffic flow prediction is essential part of transportation, traffic control, reports of accidents and intelligent transportation systems. To satisfy the demand of traffic flow prediction, this paper presents the method of real-time traffic flow prediction based on Bayesian classifier and support vector regression (SVR). We first model the traffic flow and its relations on the roads using 3D Markov random fields in spatiotemporal domain. Based on their relations, we define cliques as combination of current road and its neighbors. The dependencies on the defined cliques are estimated by using multiple linear regression and SVR. Finally, the traffic flow at next time stamp is predicted by finding the speed level with decreasing the energy function. To evaluate the performance of the proposed method, it was tested on traffic data obtained from Gyeongbu expressway. The experimental results showed that the approach using SVR-based estimation had superior accuracy than linear-based regression.

Double-smoothing for bias reduction in local linear regression

Support vector regression modeling in recursive just-in-time learning framework for adaptive soft sensing of naphtha boiling point in crude distillation unit