摘要
本文给出一种基于径向基函数(RBF)神经网络参数优化的ICA-RBF神经网络算法.首先采用帝国竞争算法(ICA)对径向基函数神经网络参数进行全局寻优,得到具有全局最优的参数初始值.然后使用梯度方法训练径向基网络模型,建立ICA-RBF神经网络算法.最后通过数值实验对该方法的有效性进行检验.
In this paper we give an ICA-RBF neural network algorithm based on optimizing the parameters of radial basis function(RBF)neural network.First,the impire competition algorithm(ICA)is used to obtain the global optimal initial values of parameters,and then by applying the gradient basis method to train the radial basis network model an ICA-RBF neural network algorithm is established.Finally,the effectiveness of the method is demonstrated through numerical experiments.
作者
李友云
谭莉慧
霍剑光
Li Youyun;Tan Lihui;Huo Jianguang(School of Mathematics and Statistics,Changsha University of Science and Technology,Changsha 410114,China)
出处
《数学理论与应用》
2019年第3期104-112,共9页
Mathematical Theory and Applications
关键词
帝国竞争算法
RBF神经网络
BP神经网
材料性能预测
Imperie competition algorithm
RBF neural network
BP neural network
Material performance prediction
