用改进RPCL法提取聚类最佳数目.docVIP

HYPERLINK /qikan/periodical.articles/shdxxb/index.html 上海大学学报JOURNAL OF SHANGHAI UNIVERSITY1999年 第5卷 第5期 Vol.5 No.5 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,如何选择适当类的数目是一个难以解决的问题.有人提出了次胜者受罚的竞争学习(rival penalized competitive learning : RPCL)算法试图来解决这一问题.但是,当数据类有重叠以及输入矢量含有非独立项时,RPCL算法的性能不能令人满意.本文提出了一种结合全协方差矩阵的RPCL算法,并逐步删除那些只包含少量训练数据的类.这种算法,我们称之为改进的RPCL算法.我们用改进的RPCL算法来确定高斯混合分布类的数目,并将其与原来的RPCL进行比较.实验证明,改进的RPCL算法比原来的RPCL算法能够更好地表征类.关键词:聚类;RPCL算法;竞争学习中图分类号:TN 912.34　　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 of Computer, Shanghai Maritime University, Shanghai 200135, China) Abstract: Selecting an appropriate number of clusters is a problem in the classical K-means algorithm. The rival penalized competitive learning(RPCL) algorithm is designed to solve this problem. But its performance is not satisfactory when the data have overlapped clusters and the input vectors contain dependent components. 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 of a 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 　　聚类是将多维空间内的数据集合分成多个有意义的子群或类的过程.它在很多领域都有着非常重要的应用,尤其是在语音、图像识别等方面.多年来,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,同时重新计算类的中心.这种方法计算简单、有效,但其缺点是需事先确定类的数目.然而在许多场合,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习(rival penalized competitive learning:RPCL)规则,来自动决定类的适当数目.它的基本思想是: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,使之远离输入值.但是,当数据类有重叠时,RPCL算法的性能