Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariser
Abstract This study proposes a novel multi‐view soft block diagonal representation framework for clustering complete and incomplete multi‐view data. First, given that the multi‐view self‐representation model offers better performance in exploring the intrinsic structure of multi‐view data, it can be...
| Huvudupphovsmän: | , , , , , |
|---|---|
| Materialtyp: | Artikel |
| Språk: | English |
| Publicerad: |
Wiley
2021-12-01
|
| Serie: | IET Computer Vision |
| Ämnen: | |
| Länkar: | https://doi.org/10.1049/cvi2.12077 |
| _version_ | 1828148956003041280 |
|---|---|
| author | Yongli Hu Cuicui Luo Boyue Wang Junbin Gao Yanfeng Sun Baocai Yin |
| author_facet | Yongli Hu Cuicui Luo Boyue Wang Junbin Gao Yanfeng Sun Baocai Yin |
| author_sort | Yongli Hu |
| collection | DOAJ |
| description | Abstract This study proposes a novel multi‐view soft block diagonal representation framework for clustering complete and incomplete multi‐view data. First, given that the multi‐view self‐representation model offers better performance in exploring the intrinsic structure of multi‐view data, it can be nicely adopted to individually construct a graph for each view. Second, since an ideal block diagonal graph is beneficial for clustering, a ‘soft’ block diagonal affinity matrix is constructed by fusing multiple previous graphs. The soft diagonal block regulariser encourages a matrix to approximately have (not exactly) K diagonal blocks, where K is the number of clusters. This strategy adds robustness to noise and outliers. Third, to handle incomplete multi‐view data, multiple indicator matrices are utilised, which can mark the position of missing elements of each view. Finally, the alternative direction of multipliers algorithm is employed to optimise the proposed model, and the corresponding algorithm complexity and convergence are also analysed. Extensive experimental results on several real‐world datasets achieve the best performance among the state‐of‐the‐art complete and incomplete clustering methods, which proves the effectiveness of the proposed methods. |
| first_indexed | 2024-04-11T21:23:02Z |
| format | Article |
| id | doaj.art-5b9c76f9d6894c9c97d4a24a507fc1c0 |
| institution | Directory Open Access Journal |
| issn | 1751-9632 1751-9640 |
| language | English |
| last_indexed | 2024-04-11T21:23:02Z |
| publishDate | 2021-12-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Computer Vision |
| spelling | doaj.art-5b9c76f9d6894c9c97d4a24a507fc1c02022-12-22T04:02:32ZengWileyIET Computer Vision1751-96321751-96402021-12-0115861863210.1049/cvi2.12077Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariserYongli Hu0Cuicui Luo1Boyue Wang2Junbin Gao3Yanfeng Sun4Baocai Yin5Faculty of Information Technology Beijing University of Technology Beijing ChinaFaculty of Information Technology Beijing University of Technology Beijing ChinaFaculty of Information Technology Beijing University of Technology Beijing ChinaThe University of Sydney Business School The University of Sydney Sydney New South Wales AustraliaFaculty of Information Technology Beijing University of Technology Beijing ChinaFaculty of Information Technology Beijing University of Technology Beijing ChinaAbstract This study proposes a novel multi‐view soft block diagonal representation framework for clustering complete and incomplete multi‐view data. First, given that the multi‐view self‐representation model offers better performance in exploring the intrinsic structure of multi‐view data, it can be nicely adopted to individually construct a graph for each view. Second, since an ideal block diagonal graph is beneficial for clustering, a ‘soft’ block diagonal affinity matrix is constructed by fusing multiple previous graphs. The soft diagonal block regulariser encourages a matrix to approximately have (not exactly) K diagonal blocks, where K is the number of clusters. This strategy adds robustness to noise and outliers. Third, to handle incomplete multi‐view data, multiple indicator matrices are utilised, which can mark the position of missing elements of each view. Finally, the alternative direction of multipliers algorithm is employed to optimise the proposed model, and the corresponding algorithm complexity and convergence are also analysed. Extensive experimental results on several real‐world datasets achieve the best performance among the state‐of‐the‐art complete and incomplete clustering methods, which proves the effectiveness of the proposed methods.https://doi.org/10.1049/cvi2.12077graph theorymatrix algebrapattern clustering |
| spellingShingle | Yongli Hu Cuicui Luo Boyue Wang Junbin Gao Yanfeng Sun Baocai Yin Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariser IET Computer Vision graph theory matrix algebra pattern clustering |
| title | Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariser |
| title_full | Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariser |
| title_fullStr | Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariser |
| title_full_unstemmed | Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariser |
| title_short | Complete/incomplete multi‐view subspace clustering via soft block‐diagonal‐induced regulariser |
| title_sort | complete incomplete multi view subspace clustering via soft block diagonal induced regulariser |
| topic | graph theory matrix algebra pattern clustering |
| url | https://doi.org/10.1049/cvi2.12077 |
| work_keys_str_mv | AT yonglihu completeincompletemultiviewsubspaceclusteringviasoftblockdiagonalinducedregulariser AT cuicuiluo completeincompletemultiviewsubspaceclusteringviasoftblockdiagonalinducedregulariser AT boyuewang completeincompletemultiviewsubspaceclusteringviasoftblockdiagonalinducedregulariser AT junbingao completeincompletemultiviewsubspaceclusteringviasoftblockdiagonalinducedregulariser AT yanfengsun completeincompletemultiviewsubspaceclusteringviasoftblockdiagonalinducedregulariser AT baocaiyin completeincompletemultiviewsubspaceclusteringviasoftblockdiagonalinducedregulariser |