Fuzzy Kernel Based Effective Clustering Techniques in Analyzing Heterogeneous Databases

Abstract

The aim of this paper is to introduce an effective fuzzy clustering technique based kernel function to find appropriate subgroups in heterogeneous databases. This paper introduces the effective fuzzy clustering that incorporates weighted bias field information, kernel distance, possibilistic memberships and fuzzy memberships into memberships equation and prototype equation. The effectiveness and efficiency of the proposed clustering techniques have been shown through the experimental results on benchmark heterogeneous databases.