Fuzzy Neural Network Classifier Based On GA And BP Algorithm And Its Application In Automatic Recognition Of ECG

With the process of time and the development of science, the inter-infiltration and cooperation among fuzzy technique, neural network technique and evolution calculation, has been an undoubted direction of development of technique in these days. As the most important and vital information-processing techniques in the intelligent information-processing field, they need more inter-combination and inter-acceleration instantly, when they are applied in the projects and to have more development. Their relation has been so close that they can't be detached clearly.This paper applies Fuzzy Neural Network based on Genetic Algorithm to the basic research of diagnosis of heart disease. Because ANN can store useful information in distributed manner, we usually make use of the topological structure and weights of neural network to realize nonlinear mapping. We combine fuzzy theory with neural network, using fuzzy inference to supervise the establish of topological structure and the initialization of weights of neural network, which make them full of meaning, also reserving the training algorithm of ANN to perfect the fuzzy logic system. Genetic algorithm has the feature of parallel and global optimization, which mimic Darwin's evolutional model of Survival-of-the-fittest. After analysis of each characters of GA and BP algorithm, against the BP neural network's deficiency of easily falling into local minimum, we propose to use a algorithm combined GA with BP to overcome BP neural network's inherent deficiency. A new fuzzy neural network classifier has been given in this dissertation. The results show that it is good and the rate of correct recognition get to what we want.