A New Cluster Validity Index for Fuzzy Clustering Based on Similarity Measure

摘要

In this paper, first, the main problems of some cluster validity indices when they have been applied to Gustafson and Kessel (GK) clustering approach are review. It is shown that most of these cluster validity indices have serious shortcomings to validate Gustafson Kessel algorithm. Then, a new cluster validity index based on a similarity measure of fuzzy clusters for validation of GK algorithm is presented. This new index is not based on a geometric distance and can determine the degree of correlation of the clusters. Finally, the proposed cluster validity index is tested and validated by using five sets of artificially generated data. The results show that the proposed cluster validity index is more efficient and realistic than the former traditional indices.