QSPR Study on the prediction of ionization potential of various organic compounds by heuristic method and radial basis function neural network

Quantitative structure-property relationship study was performed for the prediction of ionization potential (IP) of some organic compounds. Heuristic method (HM) was used to select the most appropriate molecular descriptors. Stepwise multiple linear regression (MLR) and nonlinear radial basis function neural network (RBFNN) were used to build the models. The statistical parameters provided by the MLR model were R² = 0.943; F = 953.469; RMS = 0.1797 for the training set, and R² = 0.952; F = 827.658; RMS = 0.1687 for the external test set. The RBFNN model gave better results: R² = 0.961; F = 4306.030; RMS = 0.1486 for the training set and R² = 0.955; F = 891.009; RMS = 0.1654 for test set. The predicted results were in good agreement with the experimental values.