Distributed Fuzzy Clustering with Automatic Detection of the Number of Clusters

摘要

We present a consensus-based algorithm to distributed fuzzy clustering that allows automatic estimation of the number of clusters. Also, a variant of the parallel Fuzzy c-Means algorithm that is capable of estimating the number of clusters is introduced. This variant, named DFCM, is applied for clustering data distributed across different data sites. DFCM makes use of a new, distributed version of the Xie-Beni validity criterion. Illustrative experiments show that for sites having data from different populations the developed consensus-based algorithm can provide better results than DFCM.