BE-ECM: Belief Entropy-based Evidential C-Means and its application in data clustering

Jixiang Deng, Guohui Zhou, Yong Deng, Kang Hao Cheong

Abstract: Highlights•Propose BE-ECM to address local optima problem in Evidential C-Means (ECM)•Integrate Maximum Entropy Principle (MEP) to enhance clustering stability•Design objective function with distortion and belief entropy to reduce bias•Optimize through iterative credal partition update and centroids update•Demonstrate superior clustering performance and stability in experiments