|本期目录/Table of Contents|

[1]赵永红,沈益,胡瑞芳.基于稀疏表示的球面梯度下降算法[J].浙江理工大学学报,2020,43-44(自科五):714-721.
 ZHAO Yonghong,SHEN Yi,HU Ruifang.Spherical gradient descent method based on sparse representation[J].Journal of Zhejiang Sci-Tech University,2020,43-44(自科五):714-721.
点击复制

基于稀疏表示的球面梯度下降算法()
分享到:

浙江理工大学学报[ISSN:1673-3851/CN:33-1338/TS]

卷:
第43-44卷
期数:
2020年自科五期
页码:
714-721
栏目:
出版日期:
2020-09-18

文章信息/Info

Title:
Spherical gradient descent method based on sparse representation
文章编号:
1673-3851 (2020) 05-0714-08
作者:
赵永红沈益胡瑞芳
1.浙江理工大学理学院,杭州 310018;2.嘉兴学院南湖学院数理系,浙江 嘉兴 314000
Author(s):
ZHAO Yonghong SHEN Yi HU Ruifang
1.School of Science, Zhejiang Sci-Tech University, Hangzhou 310018, China; 2. Department of Mathematics and Information Engineering, Nanhu College, Jiaxing University, Jiaxing 314001, China
关键词:
稀疏约束球面IHT1Bit压缩感知稀疏主成分分析
分类号:
TP391-41
文献标志码:
A
摘要:
针对基于稀疏表示的球面最小化问题,结合梯度下降、球面投影、稀疏逼近等方法设计了球面上的迭代硬阈值(Iterative hard thresholding,IHT)算法。首先证明了该算法产生的序列收敛到模型的L稳定点,然后通过Nesterov加速进一步提升了该算法的性能,最后将加速后的算法应用于稀疏主〖JP〗成分分析(Sparse principal component analysis, SPCA)和1Bit压缩感知(1Bit compressive sensing, 1Bit CS)。采用高斯随机矩阵进行测试,并与1Bit CS中的二进制迭代硬阈值算法、SPCA算法中的截断幂法进行了对比,数值实验表明:该算法可以有效地求解基于稀疏表示的球面最小化问题,算法产生的序列收敛到优化模型的L稳定点,加速后算法的收敛速度优于原求解算法。

参考文献/References:

[1] Elad M. Sparse and Redundant Representations: From Theory to Applications in Signal and Image Processing[M]. Springer, 2010:169-181.
[2] Zou H, Hastie T, Tibshirani R. Sparse principal component analysis[J]. Journal of Computational and Graphical Statistics, 2006, 15(2): 265-286.
[3] Boufounos P T, Baraniuk R G. 1Bit compressive sensing[C]//2008 42nd Annual Conference on Information Sciences and Systems(CISS). Princeton, NJ, USA: IEEE, 2008, 3: 16-21.
[4] Luss R, Teboulle M. Convex approximations to sparse PCA via lagrangian duality[J]. Operations Research Letters, 2011, 39(1):57-61.
[5] Luss R, Teboulle M. Conditional gradient algorithms for rankone matrix approximations with a sparsity constraint[J]. SIAM Review, 2013, 55(1): 65-98.
[6] Yuan X T, Zhang T. Truncated power method for sparse eigenvalue problems[J]. Journal of Machine Learning Research, 2013, 14: 899–925
[7] Merola G M, Chen G M. Projection sparse principal component analysis: An efficient least squares method[J]. Journal of Multivariate Analysis, 2019, 173: 366-382.
[8] Hager W W, Phan D T, Zhu J J. Projection algorithms for nonconvex minimization with application to sparse principal component analysis[J]. Journal of Global Optimization, 2016, 65(4): 657-676.
[9] Jacques L, Laska J N, Boufounos P T, et al. Robust 1bit compressive sensing via binary stable embeddings of sparse vectors[J]. IEEE Transactions on Information Theory, 2013, 59(4): 20822102.
[10] Yan M, Yang Y, Osher S. Robust 1bit compressive sensing using adaptive outlier pursuit[J]. IEEE Transactions on Signal Processing, 2012, 60(7): 3868-3875.

备注/Memo

备注/Memo:
收稿日期:2020-01-09
网络出版日期:2020-06-03
基金项目:浙江省自然科学基金杰出青年基金项目(LR19A010001);嘉兴学院南湖学院重点研究项目(N41472001-40)
作者简介:赵永红(1994-),女,安徽淮北人,硕士研究生,主要从事小波分析与压缩感知方面的研究
通信作者:胡瑞芳,E-mail:ruifanghu@qq.com
更新日期/Last Update: 2020-09-15