Simultaneous modeling of the Kovats retention indices on OV-1 and SE-54 stationary phases using artificial neural networks.

In this study, a quantitative structure-property relationship technique has been used for the simultaneous prediction of Kovats retention indices for some esters, alcohols, aldehyde and ketones on OV-1 and SE-54 stationary phases, using an artificial neural network (ANN). The best-selected descriptors that appear in the models are the molecular values, number of atoms in each molecule, molecular shadow area on the xy plane and the energy level of the highest occupied molecular orbital. A 4-6-2 ANN was generated using these descriptors as inputs and its outputs will be the Kovats retention indices on OV-1 and SE-54 stationary phases. After optimization of the network parameters, the network was trained using a training set. For the evaluation of the predictive power of the generated ANN, an optimized network was used to predict the Kovats retention indices of the prediction set. The results obtained in this study showed that the average percentage deviation between the predicted ANN and the experimental values of Kovats retention indices for the prediction set were 2.5 and 3.0% on the OV-1 and SE-54 stationary phases, respectively. These values are in good agreement with the experimental results.