Study On the Radical Basis Function Neural Network Based On Niche Genetic Algorithms

摘要

When constructing a radial basis function (RBF) neural network with clustering method, the prediction of the RBF neural network is usually influenced by the distribution of training samples. The quality of RBF neural network are difficult to catch the best. In this paper, it presents to through changing the traditional method with the niche GA to solve the matters above. Through tests, it showed that the quality of new RBF neural network built by niche genetic algorithms (NGA) performed better than the traditional one.