Low-Rank Regularized Heterogeneous Tensor Decomposition for Subspace Clustering
- 4 September 2017
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 25 (3) , 333-337
Abstract
This letter proposes a low-rank regularized heterogeneous tensor decomposition (LRRHTD) algorithm for subspace clustering, in which various constrains in different modes are incorporated to enhance the robustness of the proposed model. More specifically, due to the presence of noise and redundancy in the original tensor, LRRHTD seeks a set of orthogonal factor matrices for all but the last mode to map the high-dimensional tensor into a low-dimensional latent subspace. Furthermore, by imposing a low-rank constraint on the last mode, which is relaxed by using a nuclear norm, the lowest-rank representation that reveals the global structure of samples is obtained for the purpose of clustering. We develop an effective algorithm based on the augmented Lagrange multiplier (ALM) to optimize the proposed model. Experiments on public datasets demonstrate that our method reaches convergence within a small number of iterations and achieves promising results in comparison with state-of-the-art.Keywords
Funding Information
- National Natural Science Foundation of China (61572356, 61303208, 61701341)
This publication has 26 references indexed in Scilit:
- Nonnegative Matrix and Tensor Factorizations : An algorithmic perspectiveIEEE Signal Processing Magazine, 2014
- Sparse tensor discriminant analysisIEEE Transactions on Image Processing, 2013
- Non-negative Tri-factor tensor decomposition with applicationsKnowledge and Information Systems, 2012
- A Singular Value Thresholding Algorithm for Matrix CompletionSIAM Journal on Optimization, 2010
- Projected Gradient Methods for Nonnegative Matrix FactorizationNeural Computation, 2007
- Controlling Sparseness in Non-negative Tensor Factorization2006
- Orthogonal Tensor DecompositionsSIAM Journal on Matrix Analysis and Applications, 2001
- A Multilinear Singular Value DecompositionSIAM Journal on Matrix Analysis and Applications, 2000
- On the Best Rank-1 and Rank-(R1 ,R2 ,. . .,RN) Approximation of Higher-Order TensorsSIAM Journal on Matrix Analysis and Applications, 2000
- Some Mathematical Notes on Three-Mode Factor AnalysisPsychometrika, 1966