AN AUTOMATIC ADJUSTMENT METHOD OF BACKPROPAGATION LEARNING PARAMETERS, USING FUZZY INFERENCE

In this work, we introduce a fuzzy inference in conventional backpropa gation learning algorithm, for networks of neuron like units. This pro cedure repeatedly adjusts the learning parameters and leads the system to converge at the earliest possible time. This technique is appropri ate in a sense that optimum learning parameters are being applied in e very learning cycle automatically, whereas the conventional backpropag ation doesn't contain any well-defined rule regarding the proper deter mination of the value of learning parameters.