为了确定动态径向基函数神经网络(RBF网络)的参数以获取更优的性能,分析了影响RBF网络性能的主要因素并给出各参数的确定原则。通过实验说明确定最优RBF网络参数的方法,证明其建立的预测模型训练误差小且预测精度高。实验结果也表明,RBF网络预测精度高于BP网络,且训练时间远比BP网络的少。 In order to determine the parameters of dynamic radial basis function neural network (RBF network) for better performance, the main factors that affect the performance of RBF network are analyzed and the principle of determining each parameter is given. The method of determining the optimal RBF network parameters is demonstrated experimentally, and the training error of the established prediction model is proved to be small and the prediction accuracy is high. The experimental results also show that the prediction accuracy of RBF network is higher than that of BP network and the training time is far less than that of BP network.