New FCM-Based Algorithms for Finding the Number of Clusters

摘要：

Automatic determination of the number of dusters is a very important issue in cluster analysis. In this paper, we explore Fuzzy C-Means (FCM) baaed clustering algorithms to determine the number of dusters in a data set through cluster validity optimization. To improve the computation speed, we propose two strategies for eliminating and for spliting a cluster allowing the FCM-based algorithms to make efficient use of cluster centers computed at each step. To improve existing validity measures, we make use of a new validity function that fits particularly data sets containing overlapping clusters. Experimental results will be given to illustrate the performance of the new algorithms.

作者: Shcngrui WANG Haojun SUN Qingshan JIANG Milorad Krncta

作者单位: Dcpt. of Math. and Computer Sciences,University of Shcrbrookc,Shcrbrookc, Qc, Canada J1K 2R1 Gcncration 5 243 Consumers, 8th Floor Toronto ON. M2J 4W8

会议类型: 国际会议

会议名称: 8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)

会议地点: 上海

会议语种:英文

页码: 602-607

在线出版日期: 2001-11-14（万方平台首次上网日期，不代表论文的发表时间）