Unsupervised Kernel-Induced Fuzzy Possibilistic C-Means Technique in Investigating Real-World Data

Abstract

Abstract The goal of this study is to break down a large dataset into meaningful groupings. Due to the vast dimension and significant resemblance seen among data, exploring divided clusters in real-world datasets is the most difficult assignment. As a result, this work proposes a fuzzy set-based unsupervised effective clustering technique that includes possibilistic memberships, and fuzzy membership degrees into the membership, weighted Cauchy kernel-based similarity measure and center equations. The empirical findings demonstrate the feasibility of the proposed effective clustering technique.