QSPR model of Henry’s law constant for a diverse set of organic chemicals based on genetic algorithm-radial basis function network approach

摘要: Abstract Six quantitative structure–property relationship (QSPR) models for a diverse set of experimental data Henry’s law constant ( H ) organic chemicals under environmental condition T  = 25 °C; water–air system) have been developed based on four different molecular descriptor sets. Three the descriptors CODESSA (Comprehensive Descriptors Structural and Statistical Analysis), Tsar, Dragon software model combined from these packages, in addition HYBOT software, established using stepwise regression method. The gave best results. Furthermore, genetic algorithm was used selection descriptors, radial basis function network utilized to establish with low root mean square error (RMSE). results this study were compared well-known bond contribution group methods. method failed predict 170 all 940 compounds data-set. RMSEs 0.693, 0.798, 0.564 achieved contribution, QSPR study, respectively, logarithm . Analysis showed that hydrogen bonding between solute water as solvent has greatest influence partitioning phenomenon.