A simple filter for detecting low-rank submatrices

Research output: Contribution to journalArticle

Abstract

We present a simple algorithm for detecting low-rank submatrices from within a much larger matrix. This algorithm relies on a basic geometric property of high-dimensional space: random 2-d projections of eccentric gaussian distributions are typically concentrated in opposite quadrants of the plane.

Original languageEnglish (US)
Pages (from-to)2682-2690
Number of pages9
JournalJournal of Computational Physics
Volume231
Issue number7
DOIs
StatePublished - Apr 1 2012

Fingerprint

filters
quadrants
Gaussian distribution
eccentrics
normal density functions
projection
matrices

Keywords

  • Biclustering
  • Random projection

ASJC Scopus subject areas

  • Computer Science Applications
  • Physics and Astronomy (miscellaneous)

Cite this

A simple filter for detecting low-rank submatrices. / Rangan, Aaditya.

In: Journal of Computational Physics, Vol. 231, No. 7, 01.04.2012, p. 2682-2690.

Research output: Contribution to journalArticle

@article{56b2f16b395449eda855de80c0d167f7,
title = "A simple filter for detecting low-rank submatrices",
abstract = "We present a simple algorithm for detecting low-rank submatrices from within a much larger matrix. This algorithm relies on a basic geometric property of high-dimensional space: random 2-d projections of eccentric gaussian distributions are typically concentrated in opposite quadrants of the plane.",
keywords = "Biclustering, Random projection",
author = "Aaditya Rangan",
year = "2012",
month = "4",
day = "1",
doi = "10.1016/j.jcp.2011.12.032",
language = "English (US)",
volume = "231",
pages = "2682--2690",
journal = "Journal of Computational Physics",
issn = "0021-9991",
publisher = "Academic Press Inc.",
number = "7",

}

TY - JOUR

T1 - A simple filter for detecting low-rank submatrices

AU - Rangan, Aaditya

PY - 2012/4/1

Y1 - 2012/4/1

N2 - We present a simple algorithm for detecting low-rank submatrices from within a much larger matrix. This algorithm relies on a basic geometric property of high-dimensional space: random 2-d projections of eccentric gaussian distributions are typically concentrated in opposite quadrants of the plane.

AB - We present a simple algorithm for detecting low-rank submatrices from within a much larger matrix. This algorithm relies on a basic geometric property of high-dimensional space: random 2-d projections of eccentric gaussian distributions are typically concentrated in opposite quadrants of the plane.

KW - Biclustering

KW - Random projection

UR - http://www.scopus.com/inward/record.url?scp=84856565227&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84856565227&partnerID=8YFLogxK

U2 - 10.1016/j.jcp.2011.12.032

DO - 10.1016/j.jcp.2011.12.032

M3 - Article

AN - SCOPUS:84856565227

VL - 231

SP - 2682

EP - 2690

JO - Journal of Computational Physics

JF - Journal of Computational Physics

SN - 0021-9991

IS - 7

ER -