QSAR/QSPR/QSTR modeling and chemical grouping approach are presented to provide information on the biological properties of various substituted benzene derivatives. A novel descriptor, viz., the square of electrophilicity index (ω2 ) is proposed to provide a compact correlation between the structure of the compounds and their biological properties which is marginally superior to electrophilicity index (ω) or ω3 in most of the cases, and more or less similar to that obtained from hydrophobicity (or lipophilicity). Besides the straightforward case study, neural networks (NN) are employed to ascertain the robustness of the QSAR model obtained by implementing multiple linear regression (MLR).
Keywords: MLR; QSTR; fathead minnow; global electronic descriptor; multilayer perceptron (MLP) neural network.
© 2018 John Wiley & Sons A/S.