Monte Carlo simulation on the Stiefel manifold via polar expansion

Motivated by applications to Bayesian inference for statistical models with orthogonal matrix parameters, we present $\textit{polar expansion},$ a general approach to Monte Carlo simulation from probability distributions on the Stiefel manifold. To bypass many of the well-established challenges of s...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Jauch, Michael, Hoff, Peter D, Dunson, David B
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Jauch, Michael
Hoff, Peter D
Dunson, David B
description Motivated by applications to Bayesian inference for statistical models with orthogonal matrix parameters, we present $\textit{polar expansion},$ a general approach to Monte Carlo simulation from probability distributions on the Stiefel manifold. To bypass many of the well-established challenges of simulating from the distribution of a random orthogonal matrix $\boldsymbol{Q},$ we construct a distribution for an unconstrained random matrix $\boldsymbol{X}$ such that $\boldsymbol{Q}_X,$ the orthogonal component of the polar decomposition of $\boldsymbol{X},$ is equal in distribution to $\boldsymbol{Q}.$ The distribution of $\boldsymbol{X}$ is amenable to Markov chain Monte Carlo (MCMC) simulation using standard methods, and an approximation to the distribution of $\boldsymbol{Q}$ can be recovered from a Markov chain on the unconstrained space. When combined with modern MCMC software, polar expansion allows for routine and flexible posterior inference in models with orthogonal matrix parameters. We find that polar expansion with adaptive Hamiltonian Monte Carlo is an order of magnitude more efficient than competing MCMC approaches in a benchmark protein interaction network application. We also propose a new approach to Bayesian functional principal components analysis which we illustrate in a meteorological time series application.
doi_str_mv 10.48550/arxiv.1906.07684
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_1906_07684</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>1906_07684</sourcerecordid><originalsourceid>FETCH-LOGICAL-a674-8397ec3a63e7a7aae10c0e31dc174fa258e240419e5f3f9535993383941932733</originalsourceid><addsrcrecordid>eNotj8FqwzAQRHXpoaT5gJ6iH7AreSXLOqamTQMpPSR3szgrIpAtI7sh_fs4aWBgYJgZeIy9SpGrSmvxhuniz7m0osyFKSv1zN6_Yz8RrzGFyEff_QacfOz5rOlEfD95chR4h713MRz52SMfYsDE6TJgP87dF_bkMIy0fPiCHT4_DvVXtvvZbOv1LsPSqKwCa6gFLIEMGkSSohUE8thKoxwWuqJCCSUtaQfOatDWAsyrOYLCACzY6v_2DtEMyXeY_pobTHOHgSuW5kOL</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Monte Carlo simulation on the Stiefel manifold via polar expansion</title><source>arXiv.org</source><creator>Jauch, Michael ; Hoff, Peter D ; Dunson, David B</creator><creatorcontrib>Jauch, Michael ; Hoff, Peter D ; Dunson, David B</creatorcontrib><description>Motivated by applications to Bayesian inference for statistical models with orthogonal matrix parameters, we present $\textit{polar expansion},$ a general approach to Monte Carlo simulation from probability distributions on the Stiefel manifold. To bypass many of the well-established challenges of simulating from the distribution of a random orthogonal matrix $\boldsymbol{Q},$ we construct a distribution for an unconstrained random matrix $\boldsymbol{X}$ such that $\boldsymbol{Q}_X,$ the orthogonal component of the polar decomposition of $\boldsymbol{X},$ is equal in distribution to $\boldsymbol{Q}.$ The distribution of $\boldsymbol{X}$ is amenable to Markov chain Monte Carlo (MCMC) simulation using standard methods, and an approximation to the distribution of $\boldsymbol{Q}$ can be recovered from a Markov chain on the unconstrained space. When combined with modern MCMC software, polar expansion allows for routine and flexible posterior inference in models with orthogonal matrix parameters. We find that polar expansion with adaptive Hamiltonian Monte Carlo is an order of magnitude more efficient than competing MCMC approaches in a benchmark protein interaction network application. We also propose a new approach to Bayesian functional principal components analysis which we illustrate in a meteorological time series application.</description><identifier>DOI: 10.48550/arxiv.1906.07684</identifier><language>eng</language><subject>Statistics - Computation</subject><creationdate>2019-06</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/1906.07684$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.1906.07684$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Jauch, Michael</creatorcontrib><creatorcontrib>Hoff, Peter D</creatorcontrib><creatorcontrib>Dunson, David B</creatorcontrib><title>Monte Carlo simulation on the Stiefel manifold via polar expansion</title><description>Motivated by applications to Bayesian inference for statistical models with orthogonal matrix parameters, we present $\textit{polar expansion},$ a general approach to Monte Carlo simulation from probability distributions on the Stiefel manifold. To bypass many of the well-established challenges of simulating from the distribution of a random orthogonal matrix $\boldsymbol{Q},$ we construct a distribution for an unconstrained random matrix $\boldsymbol{X}$ such that $\boldsymbol{Q}_X,$ the orthogonal component of the polar decomposition of $\boldsymbol{X},$ is equal in distribution to $\boldsymbol{Q}.$ The distribution of $\boldsymbol{X}$ is amenable to Markov chain Monte Carlo (MCMC) simulation using standard methods, and an approximation to the distribution of $\boldsymbol{Q}$ can be recovered from a Markov chain on the unconstrained space. When combined with modern MCMC software, polar expansion allows for routine and flexible posterior inference in models with orthogonal matrix parameters. We find that polar expansion with adaptive Hamiltonian Monte Carlo is an order of magnitude more efficient than competing MCMC approaches in a benchmark protein interaction network application. We also propose a new approach to Bayesian functional principal components analysis which we illustrate in a meteorological time series application.</description><subject>Statistics - Computation</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2019</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8FqwzAQRHXpoaT5gJ6iH7AreSXLOqamTQMpPSR3szgrIpAtI7sh_fs4aWBgYJgZeIy9SpGrSmvxhuniz7m0osyFKSv1zN6_Yz8RrzGFyEff_QacfOz5rOlEfD95chR4h713MRz52SMfYsDE6TJgP87dF_bkMIy0fPiCHT4_DvVXtvvZbOv1LsPSqKwCa6gFLIEMGkSSohUE8thKoxwWuqJCCSUtaQfOatDWAsyrOYLCACzY6v_2DtEMyXeY_pobTHOHgSuW5kOL</recordid><startdate>20190618</startdate><enddate>20190618</enddate><creator>Jauch, Michael</creator><creator>Hoff, Peter D</creator><creator>Dunson, David B</creator><scope>EPD</scope><scope>GOX</scope></search><sort><creationdate>20190618</creationdate><title>Monte Carlo simulation on the Stiefel manifold via polar expansion</title><author>Jauch, Michael ; Hoff, Peter D ; Dunson, David B</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a674-8397ec3a63e7a7aae10c0e31dc174fa258e240419e5f3f9535993383941932733</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2019</creationdate><topic>Statistics - Computation</topic><toplevel>online_resources</toplevel><creatorcontrib>Jauch, Michael</creatorcontrib><creatorcontrib>Hoff, Peter D</creatorcontrib><creatorcontrib>Dunson, David B</creatorcontrib><collection>arXiv Statistics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Jauch, Michael</au><au>Hoff, Peter D</au><au>Dunson, David B</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Monte Carlo simulation on the Stiefel manifold via polar expansion</atitle><date>2019-06-18</date><risdate>2019</risdate><abstract>Motivated by applications to Bayesian inference for statistical models with orthogonal matrix parameters, we present $\textit{polar expansion},$ a general approach to Monte Carlo simulation from probability distributions on the Stiefel manifold. To bypass many of the well-established challenges of simulating from the distribution of a random orthogonal matrix $\boldsymbol{Q},$ we construct a distribution for an unconstrained random matrix $\boldsymbol{X}$ such that $\boldsymbol{Q}_X,$ the orthogonal component of the polar decomposition of $\boldsymbol{X},$ is equal in distribution to $\boldsymbol{Q}.$ The distribution of $\boldsymbol{X}$ is amenable to Markov chain Monte Carlo (MCMC) simulation using standard methods, and an approximation to the distribution of $\boldsymbol{Q}$ can be recovered from a Markov chain on the unconstrained space. When combined with modern MCMC software, polar expansion allows for routine and flexible posterior inference in models with orthogonal matrix parameters. We find that polar expansion with adaptive Hamiltonian Monte Carlo is an order of magnitude more efficient than competing MCMC approaches in a benchmark protein interaction network application. We also propose a new approach to Bayesian functional principal components analysis which we illustrate in a meteorological time series application.</abstract><doi>10.48550/arxiv.1906.07684</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.1906.07684
ispartof
issn
language eng
recordid cdi_arxiv_primary_1906_07684
source arXiv.org
subjects Statistics - Computation
title Monte Carlo simulation on the Stiefel manifold via polar expansion
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-19T14%3A14%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Monte%20Carlo%20simulation%20on%20the%20Stiefel%20manifold%20via%20polar%20expansion&rft.au=Jauch,%20Michael&rft.date=2019-06-18&rft_id=info:doi/10.48550/arxiv.1906.07684&rft_dat=%3Carxiv_GOX%3E1906_07684%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true