QSPR MODELLING OF STABILITY CONSTANTS OF METAL-THIOSEMICARBAZONE COMPLEXES USING MULTIVARIATE REGRESSION METHODS AND ARTIFICIAL NEURAL NETWORK

Abstract

In this study, the stability constants of metal-thiosemicarbazone complexes, logb₁₁ were determined by using the quantitative structure property relationship (QSPR) models. The molecular descriptors, physicochemical and quantum descriptors of complexes were generated from molecular geometric structure and semi-empirical quantum calculation PM7 and PM7/sparkle. The QSPR models were built by using the ordinary least square regression (QSPR_OLS), partial least square regression (QSPR_PLS), primary component regression (QSPR_PCR) and artificial neural network (QSPR_ANN). The best linear model QSPR_OLS (with k of 9) involves descriptors C5, xp9, electric energy, cosmo volume, N4, SsssN, cosmo area, xp10 and core-core repulsion. The QSPR_PLS, QSPR _PCR and QSPR_ANN models were developed basing on 9 varibles of the QSPR_OLS model. The quality of the QSPR models were validated by the statistical values; The QSPR_OLS: R²_train = 0.944, Q²_LOO = 0.903 and MSE = 1.035; The QSPR_PLS: R²_train = 0.929, R²_CV = 0.938 and MSE = 1.115; The QSPR_PCR: R²_train = 0.934, R²_CV = 0.9485 and MSE = 1.147. The neural network model QSPR_ANN with architecture I(9)-HL(12)-O(1) was presented also with the statistical values: R²_train = 0.9723, and R²_CV = 0.9731. The QSPR models also were evaluated externally and got good performance results with those from the experimental literature.