THE FUZZY CLUSTERING PROBLEMS AND POSSIBLE SOLUTIONS

With the widely used algorithms based on fuzzy logic, fuzzy clustering is attracting increasing attention. Nowadays there are many fuzzy clustering algorithms. One of their fundamental problems is to determine the optimal number of clusters, which has a deterministic effect on the clustering results. We can determine the optimal number of clusters with help of cluster validity indices. Cluster validity indices are used for estimating the quality of partitions produced by clustering algorithms and for determining the number of clusters in data. In this paper, factors determining the number of clusters in the existing fuzzy clustering are researched and their advantages and disadvantages are examined. The important task is to estimate the proper number of clusters in actual dataset. This paper describes a new validity index for fuzzy clustering (modified index E) and modifications improving its performance as cluster number selection criterion for fuzzy k-means. The proposed indexes are tested and validated using several data sets. The paper also presents experimental results concerning them.