Learning High-dimensional Gaussian Graphical Models under Total Positivity without Adjustment of Tuning Parameters

We consider the problem of estimating an undirected Gaussian graphical model when the underlying distribution is multivariate totally positive of order 2 (MTP2), a strong form of positive dependence. Such distributions are relevant for example for portfolio selection, since assets are usually positi...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Yuhao, Roy, Uma, Uhler, Caroline
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 Wang, Yuhao
Roy, Uma
Uhler, Caroline
description We consider the problem of estimating an undirected Gaussian graphical model when the underlying distribution is multivariate totally positive of order 2 (MTP2), a strong form of positive dependence. Such distributions are relevant for example for portfolio selection, since assets are usually positively dependent. A large body of methods have been proposed for learning undirected graphical models without the MTP2 constraint. A major limitation of these methods is that their structure recovery guarantees in the high-dimensional setting usually require a particular choice of a tuning parameter, which is unknown a priori in real world applications. We here propose a new method to estimate the underlying undirected graphical model under MTP2 and show that it is provably consistent in structure recovery without adjusting the tuning parameters. This is achieved by a constraint-based estimator that infers the structure of the underlying graphical model by testing the signs of the empirical partial correlation coefficients. We evaluate the performance of our estimator in simulations and on financial data.
doi_str_mv 10.48550/arxiv.1906.05159
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_1906_05159</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>1906_05159</sourcerecordid><originalsourceid>FETCH-LOGICAL-a679-b99951e80821ef9c289505e0a5d822a37f848ca00e8dcc370a2cfecdc01206c43</originalsourceid><addsrcrecordid>eNotj0FPwyAYQLl4MNMf4En-QCvQ0sJxWbQzqXGH3ptPoCumhQXodP_eOj295B1e8hB6oCQvBefkCcK3PedUkionnHJ5i0JrIDjrjnhvj2Om7WxctN7BhBtYYrTgcBPgNFq1qjevzRTx4rQJuPNpVQcfbbJnmy74y6bRLwlv9ecS0xpK2A-4W675AwSYTTIh3qGbAaZo7v-5Qd3Lc7fbZ-1787rbthlUtcw-pJScGkEEo2aQignJCTcEuBaMQVEPohQKCDFCK1XUBJgajNKKUEYqVRYb9PiXvU73p2BnCJf-d76_zhc_SRtXFA</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Learning High-dimensional Gaussian Graphical Models under Total Positivity without Adjustment of Tuning Parameters</title><source>arXiv.org</source><creator>Wang, Yuhao ; Roy, Uma ; Uhler, Caroline</creator><creatorcontrib>Wang, Yuhao ; Roy, Uma ; Uhler, Caroline</creatorcontrib><description>We consider the problem of estimating an undirected Gaussian graphical model when the underlying distribution is multivariate totally positive of order 2 (MTP2), a strong form of positive dependence. Such distributions are relevant for example for portfolio selection, since assets are usually positively dependent. A large body of methods have been proposed for learning undirected graphical models without the MTP2 constraint. A major limitation of these methods is that their structure recovery guarantees in the high-dimensional setting usually require a particular choice of a tuning parameter, which is unknown a priori in real world applications. We here propose a new method to estimate the underlying undirected graphical model under MTP2 and show that it is provably consistent in structure recovery without adjusting the tuning parameters. This is achieved by a constraint-based estimator that infers the structure of the underlying graphical model by testing the signs of the empirical partial correlation coefficients. We evaluate the performance of our estimator in simulations and on financial data.</description><identifier>DOI: 10.48550/arxiv.1906.05159</identifier><language>eng</language><subject>Statistics - Methodology</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>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,778,883</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/1906.05159$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.1906.05159$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Wang, Yuhao</creatorcontrib><creatorcontrib>Roy, Uma</creatorcontrib><creatorcontrib>Uhler, Caroline</creatorcontrib><title>Learning High-dimensional Gaussian Graphical Models under Total Positivity without Adjustment of Tuning Parameters</title><description>We consider the problem of estimating an undirected Gaussian graphical model when the underlying distribution is multivariate totally positive of order 2 (MTP2), a strong form of positive dependence. Such distributions are relevant for example for portfolio selection, since assets are usually positively dependent. A large body of methods have been proposed for learning undirected graphical models without the MTP2 constraint. A major limitation of these methods is that their structure recovery guarantees in the high-dimensional setting usually require a particular choice of a tuning parameter, which is unknown a priori in real world applications. We here propose a new method to estimate the underlying undirected graphical model under MTP2 and show that it is provably consistent in structure recovery without adjusting the tuning parameters. This is achieved by a constraint-based estimator that infers the structure of the underlying graphical model by testing the signs of the empirical partial correlation coefficients. We evaluate the performance of our estimator in simulations and on financial data.</description><subject>Statistics - Methodology</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2019</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj0FPwyAYQLl4MNMf4En-QCvQ0sJxWbQzqXGH3ptPoCumhQXodP_eOj295B1e8hB6oCQvBefkCcK3PedUkionnHJ5i0JrIDjrjnhvj2Om7WxctN7BhBtYYrTgcBPgNFq1qjevzRTx4rQJuPNpVQcfbbJnmy74y6bRLwlv9ecS0xpK2A-4W675AwSYTTIh3qGbAaZo7v-5Qd3Lc7fbZ-1787rbthlUtcw-pJScGkEEo2aQignJCTcEuBaMQVEPohQKCDFCK1XUBJgajNKKUEYqVRYb9PiXvU73p2BnCJf-d76_zhc_SRtXFA</recordid><startdate>20190612</startdate><enddate>20190612</enddate><creator>Wang, Yuhao</creator><creator>Roy, Uma</creator><creator>Uhler, Caroline</creator><scope>EPD</scope><scope>GOX</scope></search><sort><creationdate>20190612</creationdate><title>Learning High-dimensional Gaussian Graphical Models under Total Positivity without Adjustment of Tuning Parameters</title><author>Wang, Yuhao ; Roy, Uma ; Uhler, Caroline</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a679-b99951e80821ef9c289505e0a5d822a37f848ca00e8dcc370a2cfecdc01206c43</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2019</creationdate><topic>Statistics - Methodology</topic><toplevel>online_resources</toplevel><creatorcontrib>Wang, Yuhao</creatorcontrib><creatorcontrib>Roy, Uma</creatorcontrib><creatorcontrib>Uhler, Caroline</creatorcontrib><collection>arXiv Statistics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Wang, Yuhao</au><au>Roy, Uma</au><au>Uhler, Caroline</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Learning High-dimensional Gaussian Graphical Models under Total Positivity without Adjustment of Tuning Parameters</atitle><date>2019-06-12</date><risdate>2019</risdate><abstract>We consider the problem of estimating an undirected Gaussian graphical model when the underlying distribution is multivariate totally positive of order 2 (MTP2), a strong form of positive dependence. Such distributions are relevant for example for portfolio selection, since assets are usually positively dependent. A large body of methods have been proposed for learning undirected graphical models without the MTP2 constraint. A major limitation of these methods is that their structure recovery guarantees in the high-dimensional setting usually require a particular choice of a tuning parameter, which is unknown a priori in real world applications. We here propose a new method to estimate the underlying undirected graphical model under MTP2 and show that it is provably consistent in structure recovery without adjusting the tuning parameters. This is achieved by a constraint-based estimator that infers the structure of the underlying graphical model by testing the signs of the empirical partial correlation coefficients. We evaluate the performance of our estimator in simulations and on financial data.</abstract><doi>10.48550/arxiv.1906.05159</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.1906.05159
ispartof
issn
language eng
recordid cdi_arxiv_primary_1906_05159
source arXiv.org
subjects Statistics - Methodology
title Learning High-dimensional Gaussian Graphical Models under Total Positivity without Adjustment of Tuning Parameters
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-16T10%3A03%3A29IST&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=Learning%20High-dimensional%20Gaussian%20Graphical%20Models%20under%20Total%20Positivity%20without%20Adjustment%20of%20Tuning%20Parameters&rft.au=Wang,%20Yuhao&rft.date=2019-06-12&rft_id=info:doi/10.48550/arxiv.1906.05159&rft_dat=%3Carxiv_GOX%3E1906_05159%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