Artificial Neural Network and Topological Indices to Predict Retention Indices in Gas Chromatography

Abstract

A comparative study was undertaken to test the ability of different methods to predict the retention indices of a series of acrylates using statistical treatment as criteria of fit. In this paper, a three-layer back-propagation neural network was applied to analyze the QSAR of acrylates in gas chromatography on five different stationary phases. Nine topological indices, Wiener, Balaban, Harary, Shultz, Zagreb, and Randic of first, second, third and fourth order were calculated using a computer program in comparison with the multi-linear regression and stepwise methods. The results showed that the ANN model outperformed the MLR predictions. The training phase of the ANN model was extremely short owing to the high performance of the Levenberg-Marquardt algorithm.