Matrix completion via modified schatten 2/3-norm

EURASIP Journal on Advances in Signal Processing

Research

Chuan Li¹  Zhi Wang²  Xiaohu Luo³  Jincai Ha⁴ 

[1] Big Data and Intelligence Engineering School, Chongqing College of International Business and Economics, Chongqing, China;Big Data and Intelligence Engineering School, Chongqing College of International Business and Economics, Chongqing, China;College of Computer and Information Science, Southwest University, 400715, Chongqing, China;College of Computer and Information Science, Southwest University, 400715, Chongqing, China;School of Mathematics and Information Science, North Minzu University, Yinchuan, China;

关键词: Matrix completion;    regularization;    A new truncated schatten 2/3-norm;    Recommender system;   

DOI  :  10.1186/s13634-023-01027-w

 received in 2022-11-17, accepted in 2023-05-29,  发布年份 2023

来源: Springer

【 摘 要 】

Low-rank matrix completion is a hot topic in the field of machine learning. It is widely used in image processing, recommendation systems and subspace clustering. However, the traditional method uses the nuclear norm to approximate the rank function, which leads to only the suboptimal solution. Inspired by the closed-form formulation of L2/3\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$L_{2/3}$$\end{document} regularization, we propose a new truncated schatten 2/3-norm to approximate the rank function. Our proposed regularizer takes full account of the prior rank information and achieves a more accurate approximation of the rank function. Based on this regularizer, we propose a new low-rank matrix completion model. Meanwhile, a fast and efficient algorithm are designed to solve the proposed model. In addition, a rigorous mathematical analysis of the convergence of the proposed algorithm is provided. Finally, the superiority of our proposed model and method is investigated on synthetic data and recommender system datasets. All results show that our proposed algorithm is able to achieve comparable recovery performance while being faster and more efficient than state-of-the-art methods.

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

Fig. 5

40854_2023_491_Article_IEq4.gif

Fig. 3

Fig. 3

Fig. 4

Fig. 4

Fig. 4

Fig. 1

Fig. 2

Fig. 1

Fig. 11

Fig. 17

Fig. 10

Fig. 16

Fig. 3

Fig. 9

Fig. 2

12936_2023_4634_Article_IEq5.gif

12936_2023_4634_Article_IEq2.gif

Fig. 1

Fig. 2

Fig. 4

Fig. 2

Fig. 1

Fig. 13

Fig. 8

Fig. 1

Fig. 1

Fig. 3

Fig. 8

Fig. 1

Fig. 12

Fig. 11

Fig. 1

Fig. 10

Fig. 9

Fig. 8

Fig. 8

Fig. 1

Fig. 7

Fig. 7

12888_2023_4863_Article_IEq1.gif

12968_2023_940_Article_IEq3.gif

Fig. 3

Fig. 6

Fig. 3

Fig. 5

Fig. 4

Fig. 2

Fig. 4

Fig. 3

Fig. 3

Fig. 7

Fig. 3

Fig. 5

Fig. 1

Fig. 8

Fig. 4

Fig. 5

Fig. 2

Fig. 2

Fig. 3

Fig. 1

Fig. 3

Fig. 2

Fig. 3

Fig. 6

Fig. 1

Fig. 5

Fig. 7

Fig. 6

Fig. 6

Fig. 4

13690_2023_1130_Article_IEq62.gif

Fig. 1

13690_2023_1130_Article_IEq60.gif

13690_2023_1130_Article_IEq59.gif

Fig. 1

13690_2023_1130_Article_IEq57.gif

13690_2023_1130_Article_IEq56.gif

Fig. 4

13690_2023_1130_Article_IEq54.gif

13690_2023_1130_Article_IEq53.gif

13690_2023_1130_Article_IEq52.gif

13690_2023_1130_Article_IEq51.gif

13690_2023_1130_Article_IEq50.gif

13690_2023_1130_Article_IEq49.gif

Fig. 1

13690_2023_1130_Article_IEq47.gif

40517_2023_252_Article_IEq8.gif

Fig. 3

13690_2023_1130_Article_IEq45.gif

13690_2023_1130_Article_IEq44.gif

13690_2023_1130_Article_IEq43.gif

13690_2023_1130_Article_IEq40.gif

13690_2023_1130_Article_IEq39.gif

Fig. 2

13690_2023_1130_Article_IEq36.gif

13690_2023_1130_Article_IEq35.gif

13690_2023_1130_Article_IEq34.gif

13690_2023_1130_Article_IEq33.gif

13690_2023_1130_Article_IEq32.gif

Fig. 1

Fig. 1

13690_2023_1130_Article_IEq29.gif

13690_2023_1130_Article_IEq28.gif

13690_2023_1130_Article_IEq27.gif

13690_2023_1130_Article_IEq26.gif

13690_2023_1130_Article_IEq25.gif

Fig. 2

13690_2023_1130_Article_IEq22.gif

Fig. 2

13690_2023_1130_Article_IEq20.gif

40517_2023_252_Article_IEq151.gif

13690_2023_1130_Article_IEq19.gif

13690_2023_1130_Article_IEq18.gif

13690_2023_1130_Article_IEq17.gif

13690_2023_1130_Article_IEq16.gif

13690_2023_1130_Article_IEq6.gif

13690_2023_1130_Article_IEq5.gif

13690_2023_1130_Article_IEq4.gif

Fig. 1

40517_2023_252_Article_IEq140.gif