Enhanced Tensor Low-Rank and Sparse Representation Recovery for Incomplete Multi-View Clustering

Incomplete multi-view clustering (IMVC) has attracted remarkable attention due to the emergence of data with missing views in real applications. Recent methods attempt recover information address IMVC problem. However, they generally cannot fully explore underlying properties and correlations similarities across views. This paper proposes a novel Enhanced Tensor Low-rank Sparse Representation Recovery (ETLSRR) method, which reformulates problem as joint incomplete similarity graphs learning complete tensor representation recovery Specifically, ETLSRR learns intra-view constructs 3-way by stacking inter-view correlations. To alleviate negative influence noise, decomposes into two parts: sparse an intrinsic tensor, models noise true similarities, respectively. Both global low-rank local structured characteristics are considered, enhances discrimination matrix. Moreover, instead using convex nuclear norm, introduces generalized non-convex regularization biased approximation. Experiments on several datasets demonstrate effectiveness our method compared state-of-the-art methods.