Proceedings of the
The Nineteenth International Conference on Computational Intelligence and Security (CIS 2023)
December 1 – 4, 2023, Haikou, China
Multi-view Subspace Clustering via Simultaneous Nonconvex Low-rank Tensor Approximation and Structured Sparse Regularization
School of Mathematical Sciences, Shenzhen University, ShenZhen, China.
ABSTRACT
We propose a tensor-based non-convex multi-view subspace clustering model, where a orthogonal mapping and a tighter non-convex tensor rank approximation function are united to learn and identify the non-linearity of the global low-rank structural underlying subspace and fully consider the feature representations. A structured sparse regularization is defined to capture the sparsity of clusters at different levels. To improve clustering performance, hyper-Laplacian regularization is used to characterize local geometric features. In addition, a new algorithm is designed to optimize the proposed model. Experiments on six datasets demonstrate that the proposed method outperforms current state-of-the-art methods in various scenarios.
Keywords: Multi-view subspace clustering, Structured sparse regularization, Tensor low rank approximation.
Download PDF
School of Mathematical Sciences, Shenzhen University, ShenZhen, China.
ABSTRACT
We propose a tensor-based non-convex multi-view subspace clustering model, where a orthogonal mapping and a tighter non-convex tensor rank approximation function are united to learn and identify the non-linearity of the global low-rank structural underlying subspace and fully consider the feature representations. A structured sparse regularization is defined to capture the sparsity of clusters at different levels. To improve clustering performance, hyper-Laplacian regularization is used to characterize local geometric features. In addition, a new algorithm is designed to optimize the proposed model. Experiments on six datasets demonstrate that the proposed method outperforms current state-of-the-art methods in various scenarios.
Keywords: Multi-view subspace clustering, Structured sparse regularization, Tensor low rank approximation.
Download PDF