Principal components: A descent algorithm

Rebeca Salas-Boni, Esteban Tabak

Research output: Contribution to journalArticle

Abstract

A descent procedure is proposed for the search of low-dimensional subspaces of a high-dimensional space that satisfy an optimality criterion. Specifically, the procedure is applied to finding the subspace spanned by the first m singular components of an n-dimensional dataset. The procedure minimizes the associated cost function through a series of orthogonal transformations, each represented economically as the exponential of a skew-symmetric matrix drawn from a low-dimensional space.

Original languageEnglish (US)
Pages (from-to)162-175
Number of pages14
JournalJournal of Computational Physics
Volume267
DOIs
StatePublished - Jun 15 2014

Fingerprint

descent
Cost functions
costs
matrices

Keywords

  • Principal component analysis

ASJC Scopus subject areas

  • Computer Science Applications
  • Physics and Astronomy (miscellaneous)

Cite this

Principal components : A descent algorithm. / Salas-Boni, Rebeca; Tabak, Esteban.

In: Journal of Computational Physics, Vol. 267, 15.06.2014, p. 162-175.

Research output: Contribution to journalArticle

@article{f93fe5adbe854287b1b0fa38ac65599c,
title = "Principal components: A descent algorithm",
abstract = "A descent procedure is proposed for the search of low-dimensional subspaces of a high-dimensional space that satisfy an optimality criterion. Specifically, the procedure is applied to finding the subspace spanned by the first m singular components of an n-dimensional dataset. The procedure minimizes the associated cost function through a series of orthogonal transformations, each represented economically as the exponential of a skew-symmetric matrix drawn from a low-dimensional space.",
keywords = "Principal component analysis",
author = "Rebeca Salas-Boni and Esteban Tabak",
year = "2014",
month = "6",
day = "15",
doi = "10.1016/j.jcp.2014.02.033",
language = "English (US)",
volume = "267",
pages = "162--175",
journal = "Journal of Computational Physics",
issn = "0021-9991",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Principal components

T2 - A descent algorithm

AU - Salas-Boni, Rebeca

AU - Tabak, Esteban

PY - 2014/6/15

Y1 - 2014/6/15

N2 - A descent procedure is proposed for the search of low-dimensional subspaces of a high-dimensional space that satisfy an optimality criterion. Specifically, the procedure is applied to finding the subspace spanned by the first m singular components of an n-dimensional dataset. The procedure minimizes the associated cost function through a series of orthogonal transformations, each represented economically as the exponential of a skew-symmetric matrix drawn from a low-dimensional space.

AB - A descent procedure is proposed for the search of low-dimensional subspaces of a high-dimensional space that satisfy an optimality criterion. Specifically, the procedure is applied to finding the subspace spanned by the first m singular components of an n-dimensional dataset. The procedure minimizes the associated cost function through a series of orthogonal transformations, each represented economically as the exponential of a skew-symmetric matrix drawn from a low-dimensional space.

KW - Principal component analysis

UR - http://www.scopus.com/inward/record.url?scp=84896516582&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896516582&partnerID=8YFLogxK

U2 - 10.1016/j.jcp.2014.02.033

DO - 10.1016/j.jcp.2014.02.033

M3 - Article

VL - 267

SP - 162

EP - 175

JO - Journal of Computational Physics

JF - Journal of Computational Physics

SN - 0021-9991

ER -