On the Second Function of Cluster Validity Indexes
A cluster validity index for fuzzy partitions has two functions. One is to identify the number of clusters of an uniabeled data set The other is to assess the clustering result of an uniabeled data set. The first function has been thoroughly investigated. However, no researchers attach importance to the second function. This paper devises an experiment to investigate the second function of current cluster validity indexes associating with intra-cluster compactness and intercluster separation. Tested by various real world data sets, current cluster validity indexes are disappointing in terms of their second function.
cluster validity index intra-cluster compactness inter-cluster separation clustering accuracy fuzzy partition
Jian Mmei Chen
School of Mechanical Engineering, Hunan University OF International Economics, Changsha China, 410205
国际会议
太原
英文
274-278
2010-10-22（万方平台首次上网日期，不代表论文的发表时间）