Multi‐View Clustering Based on Intra‐View Heterogeneity and Inter‐View Compatibility

Abstract

ABSTRACT With the development of data mining technology, graph‐based multi‐view clustering methods have been widely studied. However, most of them assume that the sub‐features within the view have the same importance weight, which is not consistent with the actual situation. Therefore, this paper proposes a Multi‐view Clustering method based on Intra‐view Heterogeneity and Inter‐view Compatibility (MC‐IHIC). In this method, the sub‐features in each view are weighted, which not only considers the consistency between views, but also considers the differences between sub‐features in views. In addition, the proposed method obtains the data similarity graph by learning the local manifold structure and partitions the clusters while constructing the similarity graph by imposing the rank constraint, which overcomes the disadvantage that the traditional graph‐based clustering heavily depends on the similarity graph. Finally, the effectiveness of MC‐IHIC is verified by the comparative experiments on three multi‐view datasets.