Development of a quantitative structure-property relationship model for predicting the electrophoretic mobilities

Electrophoretic mobility (mu0) is the most important parameter governing the separation of solutes in capillary zone electrophoresis. In this paper, a new model was constructed by means of a multilayer neural network using extended delta-bar-delta (EDBD) algorithm to estimate complex property of electrophoretic mobilities of aliphatic carboxylates and amines from simpler experimental properties. The molecular weight (W), molecular volume (V), the code (+1 or -1) of acid and base and pK value were used as input parameters to predict electrophoretic mobility. The networks' architecture and the learning times were optimized. The optimum artificial neural networks (ANNs) could give excellent prediction results.