A reweighted nuclear norm minimization algorithm for low rank matrix recovery

Authors:
Yu-Fan Li;Yan-Jiao Zhang;Zheng-Hai Huang
Affiliations:
-;-;-
Venue:
Journal of Computational and Applied Mathematics
Year:
2014

Citing 10
Cited 0

Recovering sparse signals with a certain family of nonconvex penalties and DC programming

IEEE Transactions on Signal Processing
Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization

SIAM Review
Fixed point and Bregman iterative methods for matrix rank minimization

Mathematical Programming: Series A and B
Smoothing Nonlinear Conjugate Gradient Method for Image Restoration Using Nonsmooth Nonconvex Minimization

SIAM Journal on Imaging Sciences
An Unconstrained $\ell_q$ Minimization with $0q\leq1$ for Sparse Solution of Underdetermined Linear Systems

SIAM Journal on Optimization
A note on the complexity of L p minimization

Mathematical Programming: Series A and B - Special Issue on Large Scale Optimization: Analysis, Algorithms and Applications
Decoding by linear programming

IEEE Transactions on Information Theory
On the Performance of Sparse Recovery Via $\ell_p$-Minimization $(0 \leq p \leq 1)$

IEEE Transactions on Information Theory
Low-rank Matrix Recovery via Iteratively Reweighted Least Squares Minimization

SIAM Journal on Optimization
Restricted p---isometry property and its application for nonconvex compressive sensing

Advances in Computational Mathematics

Quantified Score

Hi-index	7.29

Visualization

Abstract

In this paper, we propose a reweighted nuclear norm minimization algorithm based on the weighted fixed point method (RNNM-WFP algorithm) to recover a low rank matrix, which iteratively solves an unconstrained L"2-M"p minimization problem introduced as a nonconvex smooth approximation of the low rank matrix minimization problem. We prove that any accumulation point of the sequence generated by the RNNM-WFP algorithm is a stationary point of the L"2-M"p minimization problem. Numerical experiments on randomly generated matrix completion problems indicate that the proposed algorithm has better recoverability compared to existing iteratively reweighted algorithms.