一种改进的最近邻聚类学习算法  

An Improved Closest Cluster Learning Algorithm

作　　者：李杰星[1] 章云[1] 符曦[1] 

机构地区：[1]广东工业大学自动化研究所,广州510090

出　　处：《控制理论与应用》2000年第5期735-738,共4页Control Theory & Applications

基　　金：广东省自然科学基金!(96 0 10 1)资助项目

摘　　要：提出了一种自适应模糊逻辑系统 (AFLS)的改进自适应学习算法 ,该算法将无导师学习算法与基于梯度信息的寻优学习算法相结和 。This paper brings up an improved closest cluster learning algorithm for adaptive fuzzy logical system(AFLS),which combines the non supervised algorithm and the gradient based algorithm.Besides,to determine the process of clustering,both influences of samples inputs and outputs are considered simultaneously,and a self adaptive method for selecting cluster radius is used.Two examples are presented to show the briefness,rapidity and effectiveness of the proposed algorithm.

关 键 词：模糊神经网络 非线性系统 最近邻聚类学习算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]