期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2014
卷号：3
期号：9
页码：3053-3061
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Cluster validity indexes have been used to evaluate the fitness of partitions produced by clustering algorithms. This paper presents a new validity index for fuzzy clustering called inter-cluster and intra-cluster separation (IC2S) index. Therefore, we proposed the function of disparity which combines the intra and inter-cluster separation existing between the clusters. The results of comparative study show that the proposed IC2S index has high ability in producing a good cluster number estimate. This performance is achieved by taking into consideration the existing disparity between clusters. To assess the new validation index, two data sets (Fisher's IRIS and Butterfly data set) were used and the results show that IC2S outperforms other clustering validation index for fuzzy c-means.
