A novel multilayer neural networks training algorithm that minimizes the probability of classification error

A new multilayer neural network training algorithm that minimizes the probability of classification error is proposed. The claim is made that such an algorithm possesses some clear advantages over the standard backpropagation (BP) algorithm. The convergence analysis of the proposed procedure is performed and a set of sufficient conditions which insure the convergence of the criterion sequence with probability one is given. An experimental comparison with the BP algorithm on two artificial pattern recognition problems is also given.