Match!

Randomized Matrix Decompositions Using R

Published on Jan 1, 2019in Journal of Statistical Software11.65
· DOI :10.18637/jss.v089.i11
N. Benjamin Erichson7
Estimated H-index: 7
,
Sergey Voronin1
Estimated H-index: 1
+ 1 AuthorsJ. Nathan Kutz38
Estimated H-index: 38
View in Source
Abstract
Matrix decompositions are fundamental tools in the area of applied mathematics, statistical computing, and machine learning. In particular, low-rank matrix decompositions are vital, and widely used for data analysis, dimensionality reduction, and data compression. Massive datasets, however, pose a computational challenge for traditional algorithms, placing significant constraints on both memory and processing power. Recently, the powerful concept of randomness has been introduced as a strategy to ease the computational load. The essential idea of probabilistic algorithms is to employ some amount of randomness in order to derive a smaller matrix from a high-dimensional data matrix. The smaller matrix is then used to compute the desired low-rank approximation. Such algorithms are shown to be computationally efficient for approximating matrices with low-rank structure. We present the R package rsvd, and provide a tutorial introduction to randomized matrix decompositions. Specifically, randomized routines for the singular value decomposition, (robust) principal component analysis, interpolative decomposition, and CUR decomposition are discussed. Several examples demonstrate the routines, and show the computational advantage over other methods implemented in R.
  • References (63)
  • Citations (27)
References63
Newest
#1Daniel AhfockH-Index: 3
#2William AstleH-Index: 15
Last.Sylvia RichardsonH-Index: 57
view all 3 authors...
8 Citations
#1Sergey Voronin (Tufts University)H-Index: 2
#2Per-Gunnar Martinsson (CU: University of Colorado Boulder)H-Index: 24
15 CitationsSource
#1Charles BouveyronH-Index: 13
#2Pierre LatoucheH-Index: 10
view all 3 authors...
3 Citations
#2Lionel MathelinH-Index: 2
Last.J. Nathan KutzH-Index: 38
view all 4 authors...
15 Citations
21 CitationsSource
9 Citations
#1Petros Drineas (RPI: Rensselaer Polytechnic Institute)H-Index: 41
#2Michael W. Mahoney (University of California, Berkeley)H-Index: 45
74 CitationsSource
#1N. Benjamin Erichson (St And: University of St Andrews)H-Index: 7
#2Carl Donovan (St And: University of St Andrews)H-Index: 11
35 CitationsSource
#1Joel A. TroppH-Index: 48
#2Alp YurtseverH-Index: 6
Last.Volkan CevherH-Index: 33
view all 4 authors...
17 Citations
7 Citations
Cited By27
Newest
#1N. Benjamin Erichson (St And: University of St Andrews)H-Index: 7
#2Steven L. Brunton (UW: University of Washington)H-Index: 27
Last.J. Nathan Kutz (UW: University of Washington)H-Index: 38
view all 3 authors...
19 CitationsSource
#1Mao GeH-Index: 2
#2Yong LvH-Index: 9
Last.Yubo MaH-Index: 1
view all 5 authors...
Source
#1Zhe Bai (LBNL: Lawrence Berkeley National Laboratory)H-Index: 3
Last.Steven L. BruntonH-Index: 27
view all 5 authors...
#1Yong LvH-Index: 9
#2Mao GeH-Index: 2
Last.Yubo MaH-Index: 1
view all 5 authors...
1 CitationsSource
#1Koki TsuyuzakiH-Index: 3
#2Hiroyuki Sato (Kyoto University)H-Index: 9
Last.Itoshi Nikaido (University of Tsukuba)H-Index: 2
view all 4 authors...
1 CitationsSource
#1Yingxin Lin (USYD: University of Sydney)H-Index: 3
#2Shila Ghazanfar (USYD: University of Sydney)H-Index: 6
Last.Jean Yee Hwa Yang (USYD: University of Sydney)H-Index: 9
view all 11 authors...
8 CitationsSource
Last.Rochelle BuffensteinH-Index: 43
view all 13 authors...
Source
#1Ga Wu (U of T: University of Toronto)H-Index: 3
#2Kai Luo (U of T: University of Toronto)
Last.Harold Soh (NUS: National University of Singapore)H-Index: 3
view all 4 authors...
Source
#1Saeed VatankhahH-Index: 5
#2Shuang LiuH-Index: 7
Last.Jamaledin BaniamerianH-Index: 2
view all 5 authors...
#1V. Charumathi (Indian Institute of Technology Madras)
#2M. Ramakrishna (Indian Institute of Technology Madras)H-Index: 4
Last.V. Vasudevan (Indian Institute of Technology Madras)H-Index: 8
view all 3 authors...
View next paperFinding Structure with Randomness: Probabilistic Algorithms for Constructing Approximate Matrix Decompositions