MULTIVARIATE NONLINEAR MODELING OF FLUORESCENCE DATA BY NEURAL-NETWORK WITH HIDDEN NODE PRUNING ALGORITHM

A hidden node pruning algorithm (HNPA) has been proposed as a method o f configuration optimization and training in multilayer feedforward ne twork. By this approach, a network initially bearing excessive hidden nodes is trained to sufficient precision and is pruned to the optimal size. Upon pruning, significant hidden nodes are determined by singula r value decomposition (SVD) of output matrix of hidden layer and are r etained. Weights and biases are preset intentionally in the pruned sys tem, then training continues. The method has been tested with simulate d nonlinear data and then applied to the modelling of a nonlinear fluo rescence data of a real multicomponent analytical system, and satisfac tory quantitative results have been achieved.