Cargando…

Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication)

Parameter expanded and standard expectation maximisation algorithms are described for reduced rank estimation of covariance matrices by restricted maximum likelihood, fitting the leading principal components only. Convergence behaviour of these algorithms is examined for several examples and contras...

Descripción completa

Detalles Bibliográficos
Autor principal: Meyer, Karin
Formato: Texto
Lenguaje:English
Publicado: BioMed Central 2008
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2674917/
https://www.ncbi.nlm.nih.gov/pubmed/18096112
http://dx.doi.org/10.1186/1297-9686-40-1-3
_version_ 1782166679513464832
author Meyer, Karin
author_facet Meyer, Karin
author_sort Meyer, Karin
collection PubMed
description Parameter expanded and standard expectation maximisation algorithms are described for reduced rank estimation of covariance matrices by restricted maximum likelihood, fitting the leading principal components only. Convergence behaviour of these algorithms is examined for several examples and contrasted to that of the average information algorithm, and implications for practical analyses are discussed. It is shown that expectation maximisation type algorithms are readily adapted to reduced rank estimation and converge reliably. However, as is well known for the full rank case, the convergence is linear and thus slow. Hence, these algorithms are most useful in combination with the quadratically convergent average information algorithm, in particular in the initial stages of an iterative solution scheme.
format Text
id pubmed-2674917
institution National Center for Biotechnology Information
language English
publishDate 2008
publisher BioMed Central
record_format MEDLINE/PubMed
spelling pubmed-26749172009-04-30 Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication) Meyer, Karin Genet Sel Evol Review Parameter expanded and standard expectation maximisation algorithms are described for reduced rank estimation of covariance matrices by restricted maximum likelihood, fitting the leading principal components only. Convergence behaviour of these algorithms is examined for several examples and contrasted to that of the average information algorithm, and implications for practical analyses are discussed. It is shown that expectation maximisation type algorithms are readily adapted to reduced rank estimation and converge reliably. However, as is well known for the full rank case, the convergence is linear and thus slow. Hence, these algorithms are most useful in combination with the quadratically convergent average information algorithm, in particular in the initial stages of an iterative solution scheme. BioMed Central 2008-01-15 /pmc/articles/PMC2674917/ /pubmed/18096112 http://dx.doi.org/10.1186/1297-9686-40-1-3 Text en Copyright © 2008 INRA, EDP Sciences
spellingShingle Review
Meyer, Karin
Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication)
title Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication)
title_full Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication)
title_fullStr Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication)
title_full_unstemmed Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication)
title_short Parameter expansion for estimation of reduced rank covariance matrices (Open Access publication)
title_sort parameter expansion for estimation of reduced rank covariance matrices (open access publication)
topic Review
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2674917/
https://www.ncbi.nlm.nih.gov/pubmed/18096112
http://dx.doi.org/10.1186/1297-9686-40-1-3
work_keys_str_mv AT meyerkarin parameterexpansionforestimationofreducedrankcovariancematricesopenaccesspublication