摘要
RBF神经网络隐含层节点数的确定一直以来是该网络设计成败的关键所在,文中采用K-means自组织聚类方法为隐含层节点的径向基函数确定合适的数据中心,先给出一个初始值,再慢慢调整,通过实验数据来确定最佳隐含层节点数。结果表明,如果隐含层设计得当,RBF网络可以很好地解决函数接近问题。
RBF neural network nodes in the hidden layer has been defined network design is the key to the success or failure of the paper bY K - means clustering method for the self- organization of hidden nodes RBF the identification of.suitable data centers,First gave an initial value, then slowly adjusted, through experimental data to determine the optimum number of nodes in the hidden layer. The results show that,if the hidden layer properly designed, RBF network can solve the problem closer to function.
出处
《计算机技术与发展》
2009年第1期103-105,共3页
Computer Technology and Development
基金
安徽大学人才建设项目
安徽省自然科学计划项目资助(2006KJ013A)
