Application of artificial neural networks for response surface modeling in HPLC method development

ABSTRACT

This paper discusses the usefulness of artificial neural networks [ANNs] for response surface modeling in HPLC method development. In this study, the combined effect of pH and mobile phase composition on the reversed-phase liquid chromatographic behavior of a mixture of salbuta-mol [SAL] and guaiphenesin [GUA], combination I, and a mixture of ascorbic acid [ASC], paracetamol [PAR] and guaiphenesin [GUA], combination II, was investigated. The results were compared with those produced using multiple regression [REG] analysis. To examine the respective predictive power of the regression model and the neural network model, experimental and predicted response factor values, mean of squares error [MSE], average error percentage [E[[r%], and coefficients of correlation [r] were compared. It was clear that the best networks were able to predict the experimental responses more accurately than the multiple regression analysis: