Feature-Inclusion Stochastic Search for Gaussian Graphical Models
We describe a serial algorithm called feature-inclusion stochastic search, or FINCS, that uses online estimates of edge-inclusion probabilities to guide Bayesian model determination in Gaussian graphical models. FINCS is compared to MCMC, to Metropolis-based search methods, and to the popular lasso;...
Gespeichert in:
| Veröffentlicht in: | Journal of computational and graphical statistics 2008-12, Vol.17 (4), p.790-808 |
|---|---|
| Hauptverfasser: | , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 808 |
|---|---|
| container_issue | 4 |
| container_start_page | 790 |
| container_title | Journal of computational and graphical statistics |
| container_volume | 17 |
| creator | Scott, James G Carvalho, Carlos M |
| description | We describe a serial algorithm called feature-inclusion stochastic search, or FINCS, that uses online estimates of edge-inclusion probabilities to guide Bayesian model determination in Gaussian graphical models. FINCS is compared to MCMC, to Metropolis-based search methods, and to the popular lasso; it is found to be superior along a variety of dimensions, leading to better sets of discovered models, greater speed and stability, and reasonable estimates of edge-inclusion probabilities. We illustrate FINCS on an example involving mutual-fund data, where we compare the model-averaged predictive performance of models discovered with FINCS to those discovered by competing methods. |
| doi_str_mv | 10.1198/106186008X382683 |
| format | Article |
| fullrecord | <record><control><sourceid>jstor_proqu</sourceid><recordid>TN_cdi_proquest_journals_235768156</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><jstor_id>25651228</jstor_id><sourcerecordid>25651228</sourcerecordid><originalsourceid>FETCH-LOGICAL-c457t-3bd6f8ec550039da80105e42acc4117147fd0985615503b02f7537e6f3807c653</originalsourceid><addsrcrecordid>eNp1kM1LAzEQxYMoWD_uXoTF--pMsvkonkqxtVDxUAVvIc1m6Zbtpia7SP97U1Y8CJ5m4Pfem-ERcoNwjzhWDwgClQBQH0xRodgJGSFnMqcS-WnaE86P_JxcxLgFABRjOSKTmTNdH1y-aG3Tx9q32arzdmNiV9ts5Uywm6zyIZubPsbatNk8mP2mtqbJXnzpmnhFzirTRHf9My_J--zpbfqcL1_ni-lkmduCyy5n61JUylnOAdi4NAoQuCuosbZAlFjIqoSx4gKTgq2BVjK970TFFEgrOLskd0PuPvjP3sVOb30f2nRSU8alUMhFEsEgssHHGFyl96HemXDQCPrYk_7bU7LcDpZt7Hz41VMuOFKqEn8ceN2mHnbmy4em1J05ND5UwbS2jpr9m_4NA2R1FA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>235768156</pqid></control><display><type>article</type><title>Feature-Inclusion Stochastic Search for Gaussian Graphical Models</title><source>Jstor Complete Legacy</source><source>JSTOR Mathematics & Statistics</source><creator>Scott, James G ; Carvalho, Carlos M</creator><creatorcontrib>Scott, James G ; Carvalho, Carlos M</creatorcontrib><description>We describe a serial algorithm called feature-inclusion stochastic search, or FINCS, that uses online estimates of edge-inclusion probabilities to guide Bayesian model determination in Gaussian graphical models. FINCS is compared to MCMC, to Metropolis-based search methods, and to the popular lasso; it is found to be superior along a variety of dimensions, leading to better sets of discovered models, greater speed and stability, and reasonable estimates of edge-inclusion probabilities. We illustrate FINCS on an example involving mutual-fund data, where we compare the model-averaged predictive performance of models discovered with FINCS to those discovered by competing methods.</description><identifier>ISSN: 1061-8600</identifier><identifier>EISSN: 1537-2715</identifier><identifier>DOI: 10.1198/106186008X382683</identifier><language>eng</language><publisher>Alexandria: Taylor & Francis</publisher><subject>Algorithms ; Bayesian analysis ; Bayesian model selection ; Bayesian networks ; Comparative analysis ; Covariance matrices ; Covariance selection ; Gaussian Graphical Methods ; Hyper-inverse Wishart distribution ; Lasso ; Markov chains ; Metropolis algorithm ; Metropolitan areas ; Modeling ; Parametric models ; Predictive modeling ; Probabilities ; Probability ; Separators ; Stochastic models ; Studies</subject><ispartof>Journal of computational and graphical statistics, 2008-12, Vol.17 (4), p.790-808</ispartof><rights>American Statistical Association, Institute of Mathematical Statistics, and Interface Foundation of North America 2008</rights><rights>2008 The American Statistical Association, the Institute of Mathematical Statistics, and the Interface Foundation of North America</rights><rights>Copyright American Statistical Association Dec 2008</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c457t-3bd6f8ec550039da80105e42acc4117147fd0985615503b02f7537e6f3807c653</citedby><cites>FETCH-LOGICAL-c457t-3bd6f8ec550039da80105e42acc4117147fd0985615503b02f7537e6f3807c653</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://www.jstor.org/stable/pdf/25651228$$EPDF$$P50$$Gjstor$$H</linktopdf><linktohtml>$$Uhttps://www.jstor.org/stable/25651228$$EHTML$$P50$$Gjstor$$H</linktohtml><link.rule.ids>314,776,780,799,828,27901,27902,57992,57996,58225,58229</link.rule.ids></links><search><creatorcontrib>Scott, James G</creatorcontrib><creatorcontrib>Carvalho, Carlos M</creatorcontrib><title>Feature-Inclusion Stochastic Search for Gaussian Graphical Models</title><title>Journal of computational and graphical statistics</title><description>We describe a serial algorithm called feature-inclusion stochastic search, or FINCS, that uses online estimates of edge-inclusion probabilities to guide Bayesian model determination in Gaussian graphical models. FINCS is compared to MCMC, to Metropolis-based search methods, and to the popular lasso; it is found to be superior along a variety of dimensions, leading to better sets of discovered models, greater speed and stability, and reasonable estimates of edge-inclusion probabilities. We illustrate FINCS on an example involving mutual-fund data, where we compare the model-averaged predictive performance of models discovered with FINCS to those discovered by competing methods.</description><subject>Algorithms</subject><subject>Bayesian analysis</subject><subject>Bayesian model selection</subject><subject>Bayesian networks</subject><subject>Comparative analysis</subject><subject>Covariance matrices</subject><subject>Covariance selection</subject><subject>Gaussian Graphical Methods</subject><subject>Hyper-inverse Wishart distribution</subject><subject>Lasso</subject><subject>Markov chains</subject><subject>Metropolis algorithm</subject><subject>Metropolitan areas</subject><subject>Modeling</subject><subject>Parametric models</subject><subject>Predictive modeling</subject><subject>Probabilities</subject><subject>Probability</subject><subject>Separators</subject><subject>Stochastic models</subject><subject>Studies</subject><issn>1061-8600</issn><issn>1537-2715</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2008</creationdate><recordtype>article</recordtype><recordid>eNp1kM1LAzEQxYMoWD_uXoTF--pMsvkonkqxtVDxUAVvIc1m6Zbtpia7SP97U1Y8CJ5m4Pfem-ERcoNwjzhWDwgClQBQH0xRodgJGSFnMqcS-WnaE86P_JxcxLgFABRjOSKTmTNdH1y-aG3Tx9q32arzdmNiV9ts5Uywm6zyIZubPsbatNk8mP2mtqbJXnzpmnhFzirTRHf9My_J--zpbfqcL1_ni-lkmduCyy5n61JUylnOAdi4NAoQuCuosbZAlFjIqoSx4gKTgq2BVjK970TFFEgrOLskd0PuPvjP3sVOb30f2nRSU8alUMhFEsEgssHHGFyl96HemXDQCPrYk_7bU7LcDpZt7Hz41VMuOFKqEn8ceN2mHnbmy4em1J05ND5UwbS2jpr9m_4NA2R1FA</recordid><startdate>20081201</startdate><enddate>20081201</enddate><creator>Scott, James G</creator><creator>Carvalho, Carlos M</creator><general>Taylor & Francis</general><general>JCGS Management Committee of the American Statistical Association, Institute of Mathematical Statistics, and Interface Foundation of North America</general><general>Taylor & Francis Ltd</general><scope>AAYXX</scope><scope>CITATION</scope><scope>JQ2</scope></search><sort><creationdate>20081201</creationdate><title>Feature-Inclusion Stochastic Search for Gaussian Graphical Models</title><author>Scott, James G ; Carvalho, Carlos M</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c457t-3bd6f8ec550039da80105e42acc4117147fd0985615503b02f7537e6f3807c653</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2008</creationdate><topic>Algorithms</topic><topic>Bayesian analysis</topic><topic>Bayesian model selection</topic><topic>Bayesian networks</topic><topic>Comparative analysis</topic><topic>Covariance matrices</topic><topic>Covariance selection</topic><topic>Gaussian Graphical Methods</topic><topic>Hyper-inverse Wishart distribution</topic><topic>Lasso</topic><topic>Markov chains</topic><topic>Metropolis algorithm</topic><topic>Metropolitan areas</topic><topic>Modeling</topic><topic>Parametric models</topic><topic>Predictive modeling</topic><topic>Probabilities</topic><topic>Probability</topic><topic>Separators</topic><topic>Stochastic models</topic><topic>Studies</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Scott, James G</creatorcontrib><creatorcontrib>Carvalho, Carlos M</creatorcontrib><collection>CrossRef</collection><collection>ProQuest Computer Science Collection</collection><jtitle>Journal of computational and graphical statistics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Scott, James G</au><au>Carvalho, Carlos M</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Feature-Inclusion Stochastic Search for Gaussian Graphical Models</atitle><jtitle>Journal of computational and graphical statistics</jtitle><date>2008-12-01</date><risdate>2008</risdate><volume>17</volume><issue>4</issue><spage>790</spage><epage>808</epage><pages>790-808</pages><issn>1061-8600</issn><eissn>1537-2715</eissn><abstract>We describe a serial algorithm called feature-inclusion stochastic search, or FINCS, that uses online estimates of edge-inclusion probabilities to guide Bayesian model determination in Gaussian graphical models. FINCS is compared to MCMC, to Metropolis-based search methods, and to the popular lasso; it is found to be superior along a variety of dimensions, leading to better sets of discovered models, greater speed and stability, and reasonable estimates of edge-inclusion probabilities. We illustrate FINCS on an example involving mutual-fund data, where we compare the model-averaged predictive performance of models discovered with FINCS to those discovered by competing methods.</abstract><cop>Alexandria</cop><pub>Taylor & Francis</pub><doi>10.1198/106186008X382683</doi><tpages>19</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1061-8600 |
| ispartof | Journal of computational and graphical statistics, 2008-12, Vol.17 (4), p.790-808 |
| issn | 1061-8600 1537-2715 |
| language | eng |
| recordid | cdi_proquest_journals_235768156 |
| source | Jstor Complete Legacy; JSTOR Mathematics & Statistics |
| subjects | Algorithms Bayesian analysis Bayesian model selection Bayesian networks Comparative analysis Covariance matrices Covariance selection Gaussian Graphical Methods Hyper-inverse Wishart distribution Lasso Markov chains Metropolis algorithm Metropolitan areas Modeling Parametric models Predictive modeling Probabilities Probability Separators Stochastic models Studies |
| title | Feature-Inclusion Stochastic Search for Gaussian Graphical Models |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-02T00%3A49%3A54IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-jstor_proqu&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Feature-Inclusion%20Stochastic%20Search%20for%20Gaussian%20Graphical%20Models&rft.jtitle=Journal%20of%20computational%20and%20graphical%20statistics&rft.au=Scott,%20James%20G&rft.date=2008-12-01&rft.volume=17&rft.issue=4&rft.spage=790&rft.epage=808&rft.pages=790-808&rft.issn=1061-8600&rft.eissn=1537-2715&rft_id=info:doi/10.1198/106186008X382683&rft_dat=%3Cjstor_proqu%3E25651228%3C/jstor_proqu%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=235768156&rft_id=info:pmid/&rft_jstor_id=25651228&rfr_iscdi=true |