NEURAL NETWORKS AND MODELING IN CHEMISTRY

Basic neural network architectures are described and discussed briefly. The basic equations for Hopfield, Hamming, ABAM, Kohonen and back-propagation learning schemes are given. The applicability of neural network in the domain of modelling is shown by an example of the modelling the relationship between the selectivity factor (SF) and two variables of the mobile phase (ethanol content and pH). The model obtained by the neural network is compared to the model obtained using quadratic polynomial function. Some of the problems inherent in the modelling with neural networks are discussed as well.