Stable algorithms for multiset canonical correlation analysis

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper is devoted to the construction of dynamical systems that converge to principal subspaces of multi-set canonical variates using root objective functions. With some modifications, these systems may be converted to new ones that converge to the actual canonical variates. The main important features of two algorithms that have been tested are that the first algorithm converges to the canonical variates corresponding to the canonical correlations of largest magnitudes, while the other converges to the canonical variates corresponding to the largest positive canonical correlations.

Original languageEnglish (US)
Title of host publication2009 American Control Conference, ACC 2009
Pages1280-1285
Number of pages6
DOIs
StatePublished - 2009
Event2009 American Control Conference, ACC 2009 - St. Louis, MO, United States
Duration: Jun 10 2009Jun 12 2009

Publication series

NameProceedings of the American Control Conference
ISSN (Print)0743-1619

Other

Other2009 American Control Conference, ACC 2009
Country/TerritoryUnited States
CitySt. Louis, MO
Period6/10/096/12/09

Keywords

  • Canonical correlation analysis
  • Polynomial dynamical systems
  • Root merit function

Fingerprint

Dive into the research topics of 'Stable algorithms for multiset canonical correlation analysis'. Together they form a unique fingerprint.

Cite this