A DEIM INDUCED CUR FACTORIZATION

被引:80
作者
Sorensen, D. C. [1 ]
Embree, Mark [2 ,3 ]
机构
[1] Rice Univ, Dept Computat & Appl Math, Houston, TX 77005 USA
[2] Virginia Tech, Dept Math & Computat Modeling, Acad Integrated Sci, Blacksburg, VA 24061 USA
[3] Virginia Tech, Data Analyt Div, Acad Integrated Sci, Blacksburg, VA 24061 USA
基金
美国国家科学基金会;
关键词
discrete empirical interpolation method; CUR factorization; pseudoskeleton decomposition; low-rank approximation; one-pass QR decomposition; EMPIRICAL INTERPOLATION METHOD; ALGORITHMS;
D O I
10.1137/140978430
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
We derive a CUR approximate matrix factorization based on the discrete empirical interpolation method (DEIM). For a given matrix A, such a factorization provides a low-rank approximate decomposition of the form A approximate to CUR, where C and R are subsets of the columns and rows of A, and U is constructed to make CUR a good approximation. Given a low-rank singular value decomposition A approximate to VSWT, the DEIM procedure uses V and W to select the columns and rows of A that form C and R. Through an error analysis applicable to a general class of CUR factorizations, we show that the accuracy tracks the optimal approximation error within a factor that depends on the conditioning of submatrices of V and W. For very large problems, V and W can be approximated well using an incremental QR algorithm that makes only one pass through A. Numerical examples illustrate the favorable performance of the DEIM-CUR method compared to CUR approximations based on leverage scores.
引用
收藏
页码:A1454 / A1482
页数:29
相关论文
共 28 条
[1]  
[Anonymous], 1997, NUMERICAL LINEAR ALG
[2]  
[Anonymous], 1998, SOFTWARE ENV TOOLS
[3]  
[Anonymous], 1998, Solution of large-scale eigenvalue problems with implicitly restarted Arnoldi methods, DOI DOI 10.1137/1.9780898719628
[4]   An implicitly restarted block Lanczos bidiagonalization method using Leja shifts [J].
Baglama, James ;
Reichel, Lothar .
BIT NUMERICAL MATHEMATICS, 2013, 53 (02) :285-310
[5]   Low-Rank Incremental methods for computing dominant singular subspaces [J].
Baker, C. G. ;
Gallivan, K. A. ;
Van Dooren, P. .
LINEAR ALGEBRA AND ITS APPLICATIONS, 2012, 436 (08) :2866-2888
[6]  
Baker Christopher G, 2004, THESIS
[7]   An 'empirical interpolation' method: application to efficient reduced-basis discretization of partial differential equations [J].
Barrault, M ;
Maday, Y ;
Nguyen, NC ;
Patera, AT .
COMPTES RENDUS MATHEMATIQUE, 2004, 339 (09) :667-672
[8]   Optimal CUR Matrix Decompositions [J].
Boutsidis, Christos ;
Woodruff, David P. .
STOC'14: PROCEEDINGS OF THE 46TH ANNUAL 2014 ACM SYMPOSIUM ON THEORY OF COMPUTING, 2014, :353-362
[9]   NONLINEAR MODEL REDUCTION VIA DISCRETE EMPIRICAL INTERPOLATION [J].
Chaturantabut, Saifon ;
Sorensen, Danny C. .
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2010, 32 (05) :2737-2764
[10]   On the compression of low rank matrices [J].
Cheng, H ;
Gimbutas, Z ;
Martinsson, PG ;
Rokhlin, V .
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2005, 26 (04) :1389-1404