The research of classification based on improved RBF neural network

The approximation accuracy of RBF network constructed by the incremental learning algorithm to the target was not high. For function approximation or other requirements of high accuracy, such accuracy of RBF network model can not meet the requirements. We have improved this network model focused on three aspects to improve the bottleneck, and have an experiment and comparatively analyze these improvements algorithm on an UCI database, the experimental results show that the improved algorithm has better performances.