Multicomponent acid-base titration by principal component-artificial neural network calibration

In this study, a three-layered feed-forward artificial neural network (ANN) trained by back-propagation learning was used to model the complex non-linear relationship between the concentration of anthranilic acid (HA), nicotinic acid (HN), picolinic acid (HP) and sulfanilic acid (HS) in their quaternary mixtures and the pH of solutions at different volumes of the added titrant. The principal components of the pH matrix were used as the input of the network. The network architecture and parameters were optimized to give low prediction error. The optimized networks predicted the concentrations of acids in synthetic mixtures. The results showed that the ANN used can proceed the titration data with low percent relative error of prediction (i.e.<4%). A comparison between the ANN and PLS methods revealed the superiority of the results obtained by the former method.