Prediction of Setschenow constants of organic compounds based on a 3D structure representation

Quantitative structure–property relationship (QSPR) studies were performed between three-dimensional (3D) descriptors representing the molecular structures and Setschenow constants (Ksalt) by sodium chloride for a diverse set of organic compounds. The entire set of 101 compounds was divided into a training set of 71 compounds and a test set of 30 compounds according to Kennard and Stones algorithm. Multilinear regression (MLR) analysis was used to select the best subset of descriptors and to build linear models; while nonlinear models were developed by means of artificial neural network (ANN). The obtained models with five descriptors involved show a good predictive power: for the test set, a squared correlation coefficient (R2) of 0.8987 and a standard error of estimation (s) of 0.021 were achieved by the MLR analysis; while by the ANN, R2 and s were 0.9034 and 0.020, respectively.