CCIM-SLR: Incomplete multiview co-clustering by sparse low-rank representation

Clustering incomplete multiview data in real-world applications has become a topic of recent interest. However, producing clustering results from multiview data with missing views and different degrees of missing data points is a challenging task. To address this issue, we propose a co-clustering method for incomplete multiview data by sparse low-rank representation (CCIM-SLR). The proposed method integrates the global and local structures of incomplete multiview data and effectively captures the correlations between samples in a view, as well as between different views by using sparse low-rank learning. CCIM-SLR can alternate between learning the shared hidden view, visible view, and cluster partitions within a co-learning framework. An iterative algorithm with guaranteed convergence is used to optimize the proposed objective function. Compared with other baseline models, CCIM-SLR achieved the best performance in the comprehensive experiments on the five benchmark datasets, particularly on those with varying degrees of incompleteness.