EXTENDED KALMAN FILTERING TRAINED NEURAL NETWORKS AND ITS APPLICATIONS TO MULTICOMPONENT SPECTRAL RESOLUTION

Backpropagation (BP), one of the most useful algorithms to train neura l networks (NN), has some deficiencies and/or inadequacies such as low convergence and local optima. A novel learning method for training NN ,extended Kalman filtering(EF), has been developed with rapid converge nce speed, few iteration cycles and small hidden neurons. This EFNN me thod was used for multicomponent spectral resolution and simultaneous determination of composite pharmaceutical APC preparations and mixed a romatic species samples with satisfactory results.