Prediction of Volatile Components Retention Time in Blackstrap Molasses by Least‐Squares Support Vector Machine

Abstract

House flies are pestiferous insects that have the potential to spread many diseases to humans and livestock, so it is very significant for us to manage house fly populations. Many commercial types of bait are available to attract house flies, but most are designed for outdoor or limited indoor use, due to their malodorous components. This study sought to identify compounds present in blackstrap molasses that might be attractive to house flies. An effective Quantitative Structure-Property Relationship (QSPR) model between the Retention Time (RT) and five molecular descriptors of the volatile compounds in blackstrap molasses, was built using a modified algorithm of Least-Squares Support Vector Machine (LS-SVM). Descriptors calculated from the molecular structures alone were used to represent the characteristics of compounds. The five molecular descriptors selected by the Heuristic Method (HM) in CODESSA were used as inputs for LS-SVM

Modeling and prediction of retention behavior of histidine‐containing peptides in immobilized metal‐affinity chromatography

“…To simplify traditional SVM, Suykens and Vandewalle [33] proposed LS-SVM. LS-SVM encompasses advantages of SVM, except that it only requires solving a set of linear equations (linear programming) [34]. In this study, the radial basis function was used as the kernel of LS-SVM.…”

Current Mathematical Methods Used in QSAR/QSPR Studies

“…Therefore, LS-SVM is faster than traditional SVM in treating the same work. The related literature is presented in [ 37 , 114 – 118 ].…”

Prediction of the binary surface tension of mixtures containing ionic liquids using Support Vector Machine algorithms

“…Owing to its virtues and outstanding generalization performance over other methods, the SVM has attracted attention and been applied widely in pattern recognition, classification and regression problems [52][53][54][55][56]. SVM has a solid theoretical foundation based on the statistical learning theory [57].…”

“…The structural risk minimization (SRM) principle is employed in SVMs instead of the traditional empirical risk minimization principle (ERM) used in classical methods (such as conventional neural network), which makes SVMs quite resistant to the under-fitting and over-fitting problems [58,59]. In addition, because of their specific formulation, sparse solutions can be obtained, and both linear and nonlinear regression can be carried out [53]. A more advantageous SVM strategy over the traditional methods based on the ANNs can be found in the literature [60,61].…”