Non-negative matrix factorisation for network reordering

Clare Lee, Desmond Higham, D. Crowther, J. Keith Vass

Research output: Contribution to journalArticle

Abstract

Non-negative matrix factorisation covers a variety of algorithms that attempt to
represent a given, large, data matrix as a sum of low rank matrices with a prescribed sign pattern. There are intuitave advantages to this approach, but also theoretical and computational challenges. In this exploratory paper we investigate the use of non-negative matrix factorisation algorithms as a means to reorder the nodes in a large network. This gives a set of alternatives to the more traditional approach of using the singular value decomposition. We describe and implement a range of recently proposed algorithms and evaluate their performance on synthetically constructed test data and on a real data set arising in cancer research.
Original languageEnglish
Pages (from-to)39-53
Number of pages16
JournalMonografias de la Real Academia de Ciencias de Zaragoza
Volume33
Publication statusPublished - 2010

Fingerprint

Factorization
Singular value decomposition

Keywords

  • matrix factorisation
  • data matrix
  • cancer research

Cite this

Lee, Clare ; Higham, Desmond ; Crowther, D. ; Vass, J. Keith. / Non-negative matrix factorisation for network reordering. In: Monografias de la Real Academia de Ciencias de Zaragoza. 2010 ; Vol. 33. pp. 39-53.
@article{077a34d442e74f9abb4e35fb316a2b73,
title = "Non-negative matrix factorisation for network reordering",
abstract = "Non-negative matrix factorisation covers a variety of algorithms that attempt to represent a given, large, data matrix as a sum of low rank matrices with a prescribed sign pattern. There are intuitave advantages to this approach, but also theoretical and computational challenges. In this exploratory paper we investigate the use of non-negative matrix factorisation algorithms as a means to reorder the nodes in a large network. This gives a set of alternatives to the more traditional approach of using the singular value decomposition. We describe and implement a range of recently proposed algorithms and evaluate their performance on synthetically constructed test data and on a real data set arising in cancer research.",
keywords = "matrix factorisation , data matrix , cancer research",
author = "Clare Lee and Desmond Higham and D. Crowther and Vass, {J. Keith}",
note = "Special issue honoring Manuel Calvo on his 65th birthday",
year = "2010",
language = "English",
volume = "33",
pages = "39--53",
journal = "Monografias de la Real Academia de Ciencias de Zaragoza",

}

Non-negative matrix factorisation for network reordering. / Lee, Clare; Higham, Desmond; Crowther, D.; Vass, J. Keith.

In: Monografias de la Real Academia de Ciencias de Zaragoza, Vol. 33, 2010, p. 39-53.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Non-negative matrix factorisation for network reordering

AU - Lee, Clare

AU - Higham, Desmond

AU - Crowther, D.

AU - Vass, J. Keith

N1 - Special issue honoring Manuel Calvo on his 65th birthday

PY - 2010

Y1 - 2010

N2 - Non-negative matrix factorisation covers a variety of algorithms that attempt to represent a given, large, data matrix as a sum of low rank matrices with a prescribed sign pattern. There are intuitave advantages to this approach, but also theoretical and computational challenges. In this exploratory paper we investigate the use of non-negative matrix factorisation algorithms as a means to reorder the nodes in a large network. This gives a set of alternatives to the more traditional approach of using the singular value decomposition. We describe and implement a range of recently proposed algorithms and evaluate their performance on synthetically constructed test data and on a real data set arising in cancer research.

AB - Non-negative matrix factorisation covers a variety of algorithms that attempt to represent a given, large, data matrix as a sum of low rank matrices with a prescribed sign pattern. There are intuitave advantages to this approach, but also theoretical and computational challenges. In this exploratory paper we investigate the use of non-negative matrix factorisation algorithms as a means to reorder the nodes in a large network. This gives a set of alternatives to the more traditional approach of using the singular value decomposition. We describe and implement a range of recently proposed algorithms and evaluate their performance on synthetically constructed test data and on a real data set arising in cancer research.

KW - matrix factorisation

KW - data matrix

KW - cancer research

UR - http://www.unizar.es/acz/05Publicaciones/Monografias/MonografiasPublicadas/Monografia33/039.pdf

M3 - Article

VL - 33

SP - 39

EP - 53

JO - Monografias de la Real Academia de Ciencias de Zaragoza

JF - Monografias de la Real Academia de Ciencias de Zaragoza

ER -