文档介绍：..页眉.. 页脚.. 上海大学学报 JOURNAL OF SHANGHAI UNIVERSITY 1999 年第5卷第5期 1999 用改进的 RPCL 算法提取聚类的最佳数目李昕郑宇江芳泽摘要:对于传统的 K平均算法来说,(rival petitive learning :RPCL) ,当数据类有重叠以及输入矢量含有非独立项时,RPC RPCL 算法,,我们称之为改进的 RPCL RPCL 算法来确定高斯混合分布类的数目,并将其与原来的 RPCL ,改进的 RPCL 算法比原来的 RPC L 算法能够更好地表征类. 关键词:聚类;RPCL 算法;竞争学****中图分类号:TN 文献标识码:A An Improved RPCL Algorithm for Clustering LI Xin JIANG Fang-ze (School of Automation, Shanghai University, Shanghai 200072, China); ZHENG Yu (Department puter, Shanghai Maritime University, Shanghai 200135, China) ..页眉.. 页脚.. Abstract: Selecting an appropriate number of clusters isa problem in the classical K-means algorithm. The rival petitive learning(RPCL) algorithm is designed to solve this problem. But its performance isnot satisfactory when the data have overlapped clusters and the input vectors contain ponents. This paper addresses this problem by incorporating full covariance matrices into the original RPCL algorithm. The resulting algorithm, referred to as the improved RPCL algorithm, progressively eliminates the units whose clusters contain only a small portion of the training data. The improved algorithm is applied to determine the number of clusters ofa Gaussian mixture distribution. The results show that the covariance matrices in the improved RPCL algorithm have a better representation of the clusters than those of the original RPCL algorithm. Key words: clustering; RPCL algorithm; competition learning ,尤其是在语音、,人们提出过许多种聚类的方法[1],其中最著名的是 K平均法,它在每次迭代过程中将对象归入相距中心最近的一类,、有效,,这个数目是未知的. 为了解决这一问题,Xu 和Krzyzak [2]提出了一种称为次胜者受罚的竞争学****rival petitive learning:RPCL) 规则,:对每个输入而言,不仅竞争获胜单元的权值被修正以适应输入值,而且对次胜单元采用惩罚的方法,,当数据类有重叠时,RPCL , RPCL 算法用欧氏距离(Euclidean distance) 来作为距