LEARNING BY FUZZIFIED NEURAL NETWORKS

We derive a general learning algorithm for training a fuzzified feedfo rward neural network that has fuzzy inputs, fuzzy targets, and fuzzy c onnection weights. The derived algorithm is applicable to the learning of fuzzy connection weights with various shapes such as triangular an d trapezoid. First we briefly describe how a feedforward neural networ k can be fuzzified. inputs, targets, and connection weights in the fuz zified neural network can be fuzzy numbers. Next we define a cost func tion that measures the difference between a fuzzy target vector and an actual fuzzy output vector Then we derive a learning algorithm from t he cost function for adjusting fuzzy connection weights. Finally we sh ow some results of computer simulations.