基于OCA客观聚类的RBF神经网络的研究  

Research on RBF Neural Network Based on OCA Objective Clustering

作　　者：黄静[1] 张江[1] 

机构地区：[1]浙江理工大学信息学院,杭州310018

出　　处：《浙江理工大学学报（自然科学版）》2014年第1期71-74,共4页Journal of Zhejiang Sci-Tech University(Natural Sciences)

基　　金：浙江省自然科学基金资助项目(LY12F03012)

摘　　要：影响RBF神经网络性能的关键因素是基函数中心的选取,而目前尚没有可靠的方法选取RBF神经网络的中心。基于GMDH理论的OCA客观聚类具有能够自动确定最优聚类个数的优点。将OCA聚类应用于RBF神经网络中,用以自适应确定隐节点数目和各径向基函数中心,克服了传统RBF网络不能客观确定隐节点数目的缺点。实验仿真结果表明,基于OCA客观聚类的RBF神经网络具有自适应性、正确率高和训练速度快的优点。The key factor influencing RBF neural network performance is the selection of basis function center. Currently, there is no reliable method for selecting the center of RBF neural network. OCA objec- tive clustering based on GMDH theory has the advantage of automatically determining the optimal cluste- ring number. This research overcomes the disadvantage of traditional RBF network that it cannot objec- tively determine the number of hidden nodes by using OCA clustering in RBF neural network to determine the number of hidden nodes and the center of each radial basis function. The result of experimental simula- tion shows that RBF neural network based on OCA objective clustering has such advantages as adaptivity, high accuracy and fast training speed.

关 键 词：RBF神经网络 OCA客观聚类 隐节点数目 基函数中心 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]