Efficient One-Pass Multi-View Subspace Clustering with Consensus Anchors

Multi-view subspace clustering (MVSC) optimally integrates multiple graph structure information to improve performance. Recently, many anchor-based variants are proposed reduce the computational complexity of MVSC. Though achieving considerable acceleration, we observe that most them adopt fixed anchor points separating from subsequential construction, which may adversely affect In addition, post-processing is required generate discrete labels with additional time consumption. To address these issues, propose a scalable and parameter-free MVSC method directly output optimal graph, termed as Efficient One-pass Subspace Clustering Consensus Anchors (EOMSC-CA). Specially, combine learning construction into uniform framework boost Meanwhile, by imposing connectivity constraint, our algorithm outputs without any procedures previous methods do. Our EOMSC-CA proven be linear respecting data size. The superiority over effectiveness efficiency demonstrated extensive experiments. code publicly available at https://github.com/Tracesource/EOMSC-CA.