Multi-view K-means clustering with bregman divergences

Abstract

Multi-view clustering has become an important task in machine learning. How to make full use of the similarities and differences among multiple views to generate clusters is a crucial issue. However, the existing multi-view clustering methods rarely consider the redundancy of the multiple views. In this paper, we propose a novel multi-view clustering method with Bregman divergences (MVBDC), where the clustering result is achieved by minimizing the Bregman divergences between clustering results obtained by weighted multiple views and the item that controls redundancy of multiple views. The experimental results on nine data sets demonstrate that our algorithm has a good clustering performance.