Multi-view collaborative locally adaptive clustering with Minkowski metric

Recently, many heterogeneous but related views of data have been generated in a number of applications. Different views may represent distinct aspects of the same data, which often have the same or consensus cluster structure. Discovering cluster structure in multi-view data has become a hot research topic and significant progress has been made in multi-view clustering. However, it remains a challenging issue to exploit the diversity within each view and investigate the relationship across multiple views simultaneously. To address the above issues, in this paper, we extend locally adaptive clustering into a multi-view framework with Minkowski metric and propose a novel approach termed multi-view collaborative locally adaptive clustering with Minkowski metric (MV-CoMLAC). Different from the existing multi-view subspace clustering methods, the proposed approach is capable of simultaneously taking into account the subspace diversity within each view as well as the knowledge across different views. A collaborative strategy is designed to exploit the complementary information from different low-dimensional subspaces. Furthermore, Minkowski metric is utilized to take into account the influence of the L-p distance (p ≥ 0), making our method adaptive to different application tasks. Extensive experiments have been conducted on several multi-view datasets, which demonstrate the superiority of our approach over the existing multi-view clustering methods.