TW-Co-MFC: Two-Level Weighted Collaborative Multi-view Fuzzy Clustering Based on Maximum Entropy

Abstract

In recent years, multi-view clustering research has attracted more and more attention because of the rapidly growing of unsupervised analysis demand towards multi-view data in practical applications. Despite the significant advances in multi-view clustering, there are still two challenges facing us: (1) how to make full use of the consistency and complementary information in multiple views and (2) how to discriminate the contributions of different views and features in the same view to reveal efficiently the latent cluster structure of multi-view data for clustering. In this paper, we propose a novel two-level weighted collaborative multi-view fuzzy clustering approach (TW-Co-MFC) to address the above issues. In TW-Co-MFC, a two-level weighting strategy is devised to measure the importance of views and features, and a collaborative working mechanism is introduced to balance the within-view clustering quality and the cross-view clustering consistency, then an iterative optimization objective function based on the maximum entropy principle is designed for multi-view clustering. Experiments on real-world datasets show the effectiveness of the proposed approach.