A BP Neural Network Diagnosis Based on Fuzzy Classification

The paper proposed an new diagnosis algorithm of back-propagation combine with fuzzy classify approach. For some complexity environments, the classic BP neural network appear limitation on classify issues. The paper imposed fuzzy model on neural network structure, which using classify variance and energy function to adjust convergence of the network. With the improved nonlinear mapping property, the diagnostical processing, in our experiments, shows perfect results with identify ratio of 100%, while classical experiment is 50% accordingly.