The Neural Network with Fuzzy Inference Function

摘要

Based on structure and algorithm of BP neural network and the characteristic of fuzzy inference system, the paper presents one new model of BP neural network which is merged with the function of fuzzy reasoning. In the model, the power of fuzzy reasoning is merged into the neuron of hidden layer so that the fuzzy neural network could excite fuzzy reasoning automatically according to the system input. At the same time, it could finish the nonlinear mapping using the BP algorithm. In the paper, we give some detailed algorithms of fuzzy rules extraction and fuzzy neural network.