论文部分内容阅读
BP网络是广为应用的一种前馈网络,但是由于其自身缺陷,其它网络越来越受到关注.目前RBF网络是前馈网络研究中的一个热点,有关其逼近理论以及算法的研究出现在许多文献之中.本文综述了RBF网络理论,对其结构、函数逼近性质、学习算法做了较为详细的介绍,并指出了RBF网络各种学习算法的优点和存在的问题.
BP network is widely used as a feedforward network, but due to its own defects, other networks are getting more and more attention.Nowadays, RBF network is a hot point in the research of feedforward networks, and the research about its approximation theory and algorithm appears in Many documents have been written in this paper.In this paper, we review the theory of RBF networks, introduce the structure, function approximation properties and learning algorithms in detail, and point out the advantages and problems of RBF network learning algorithms.