Sensitivity analysis for probability assessments in Bayesian networks
When eliciting a probability model from experts, knowledge engineers may compare the results of the model with expert judgment on test scenarios, then adjust model parameters to bring the behavior of the model more in line with the experts intuition. This paper presents a methodology for analytic co...
Gespeichert in:
| Veröffentlicht in: | IEEE transactions on systems, man, and cybernetics man, and cybernetics, 1995-06, Vol.25 (6), p.901-909 |
|---|---|
| 1. Verfasser: | |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 909 |
|---|---|
| container_issue | 6 |
| container_start_page | 901 |
| container_title | IEEE transactions on systems, man, and cybernetics |
| container_volume | 25 |
| creator | Laskey, K.B. |
| description | When eliciting a probability model from experts, knowledge engineers may compare the results of the model with expert judgment on test scenarios, then adjust model parameters to bring the behavior of the model more in line with the experts intuition. This paper presents a methodology for analytic computation of sensitivity values in Bayesian network models. Sensitivity values are partial derivatives of output probabilities with respect to parameters being varied in the sensitivity analysis. They measure the impact of small changes in a network parameter on a target probability value or distribution. Sensitivity values can be used to focus knowledge elicitation effort on those parameters having the most impact on outputs of concern. Analytic sensitivity values are computed for an example and compared to sensitivity analysis by direct variation of parameters.< > |
| doi_str_mv | 10.1109/21.384252 |
| format | Article |
| fullrecord | <record><control><sourceid>pascalfrancis_RIE</sourceid><recordid>TN_cdi_pascalfrancis_primary_3533476</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>384252</ieee_id><sourcerecordid>3533476</sourcerecordid><originalsourceid>FETCH-LOGICAL-c275t-da7b1499fdaacf1a7faeede6db955680be2c8dff1e61f33a359afd2a636f65953</originalsourceid><addsrcrecordid>eNpFj71LxEAUxBdRMJ4WtlYpbCxy7kd2ky31OD_gwEKtw0vyFlZzm7AvKPnvjUa0Gob5zcAwdi74Wghur6VYqzKXWh6wRApTZtJye8gSzkWZ2byQx-yE6G22eW51wrbPGMiP_sOPUwoBuok8pa6P6RD7Gmrf_QRESLTHMFLqQ3oLE5KHkAYcP_v4TqfsyEFHeParK_Z6t33ZPGS7p_vHzc0ua2Shx6yFoha5ta4FaJyAwgFii6atrdam5DXKpmydE2iEUwqUtuBaCUYZZ7TVasWult0m9kQRXTVEv4c4VYJX3_8rKarl_8xeLuwA1EDnIoTG019BaaXywszYxYJ5RPxPl40vPE1kkw</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Sensitivity analysis for probability assessments in Bayesian networks</title><source>IEEE Xplore</source><creator>Laskey, K.B.</creator><creatorcontrib>Laskey, K.B.</creatorcontrib><description>When eliciting a probability model from experts, knowledge engineers may compare the results of the model with expert judgment on test scenarios, then adjust model parameters to bring the behavior of the model more in line with the experts intuition. This paper presents a methodology for analytic computation of sensitivity values in Bayesian network models. Sensitivity values are partial derivatives of output probabilities with respect to parameters being varied in the sensitivity analysis. They measure the impact of small changes in a network parameter on a target probability value or distribution. Sensitivity values can be used to focus knowledge elicitation effort on those parameters having the most impact on outputs of concern. Analytic sensitivity values are computed for an example and compared to sensitivity analysis by direct variation of parameters.< ></description><identifier>ISSN: 0018-9472</identifier><identifier>EISSN: 2168-2909</identifier><identifier>DOI: 10.1109/21.384252</identifier><identifier>CODEN: ISYMAW</identifier><language>eng</language><publisher>New York, NY: IEEE</publisher><subject>Applied sciences ; Artificial intelligence ; Bayesian methods ; Computer science; control theory; systems ; Exact sciences and technology ; Expert systems ; Intelligent networks ; Knowledge engineering ; Learning and adaptive systems ; Network topology ; Random variables ; Sensitivity analysis ; System testing ; Uncertainty</subject><ispartof>IEEE transactions on systems, man, and cybernetics, 1995-06, Vol.25 (6), p.901-909</ispartof><rights>1995 INIST-CNRS</rights><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c275t-da7b1499fdaacf1a7faeede6db955680be2c8dff1e61f33a359afd2a636f65953</citedby><cites>FETCH-LOGICAL-c275t-da7b1499fdaacf1a7faeede6db955680be2c8dff1e61f33a359afd2a636f65953</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/384252$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,777,781,793,27905,27906,54739</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/384252$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=3533476$$DView record in Pascal Francis$$Hfree_for_read</backlink></links><search><creatorcontrib>Laskey, K.B.</creatorcontrib><title>Sensitivity analysis for probability assessments in Bayesian networks</title><title>IEEE transactions on systems, man, and cybernetics</title><addtitle>T-SMC</addtitle><description>When eliciting a probability model from experts, knowledge engineers may compare the results of the model with expert judgment on test scenarios, then adjust model parameters to bring the behavior of the model more in line with the experts intuition. This paper presents a methodology for analytic computation of sensitivity values in Bayesian network models. Sensitivity values are partial derivatives of output probabilities with respect to parameters being varied in the sensitivity analysis. They measure the impact of small changes in a network parameter on a target probability value or distribution. Sensitivity values can be used to focus knowledge elicitation effort on those parameters having the most impact on outputs of concern. Analytic sensitivity values are computed for an example and compared to sensitivity analysis by direct variation of parameters.< ></description><subject>Applied sciences</subject><subject>Artificial intelligence</subject><subject>Bayesian methods</subject><subject>Computer science; control theory; systems</subject><subject>Exact sciences and technology</subject><subject>Expert systems</subject><subject>Intelligent networks</subject><subject>Knowledge engineering</subject><subject>Learning and adaptive systems</subject><subject>Network topology</subject><subject>Random variables</subject><subject>Sensitivity analysis</subject><subject>System testing</subject><subject>Uncertainty</subject><issn>0018-9472</issn><issn>2168-2909</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>1995</creationdate><recordtype>article</recordtype><recordid>eNpFj71LxEAUxBdRMJ4WtlYpbCxy7kd2ky31OD_gwEKtw0vyFlZzm7AvKPnvjUa0Gob5zcAwdi74Wghur6VYqzKXWh6wRApTZtJye8gSzkWZ2byQx-yE6G22eW51wrbPGMiP_sOPUwoBuok8pa6P6RD7Gmrf_QRESLTHMFLqQ3oLE5KHkAYcP_v4TqfsyEFHeParK_Z6t33ZPGS7p_vHzc0ua2Shx6yFoha5ta4FaJyAwgFii6atrdam5DXKpmydE2iEUwqUtuBaCUYZZ7TVasWult0m9kQRXTVEv4c4VYJX3_8rKarl_8xeLuwA1EDnIoTG019BaaXywszYxYJ5RPxPl40vPE1kkw</recordid><startdate>19950601</startdate><enddate>19950601</enddate><creator>Laskey, K.B.</creator><general>IEEE</general><general>Institute of Electrical and Electronics Engineers</general><scope>IQODW</scope><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>19950601</creationdate><title>Sensitivity analysis for probability assessments in Bayesian networks</title><author>Laskey, K.B.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c275t-da7b1499fdaacf1a7faeede6db955680be2c8dff1e61f33a359afd2a636f65953</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>1995</creationdate><topic>Applied sciences</topic><topic>Artificial intelligence</topic><topic>Bayesian methods</topic><topic>Computer science; control theory; systems</topic><topic>Exact sciences and technology</topic><topic>Expert systems</topic><topic>Intelligent networks</topic><topic>Knowledge engineering</topic><topic>Learning and adaptive systems</topic><topic>Network topology</topic><topic>Random variables</topic><topic>Sensitivity analysis</topic><topic>System testing</topic><topic>Uncertainty</topic><toplevel>online_resources</toplevel><creatorcontrib>Laskey, K.B.</creatorcontrib><collection>Pascal-Francis</collection><collection>CrossRef</collection><jtitle>IEEE transactions on systems, man, and cybernetics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Laskey, K.B.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Sensitivity analysis for probability assessments in Bayesian networks</atitle><jtitle>IEEE transactions on systems, man, and cybernetics</jtitle><stitle>T-SMC</stitle><date>1995-06-01</date><risdate>1995</risdate><volume>25</volume><issue>6</issue><spage>901</spage><epage>909</epage><pages>901-909</pages><issn>0018-9472</issn><eissn>2168-2909</eissn><coden>ISYMAW</coden><abstract>When eliciting a probability model from experts, knowledge engineers may compare the results of the model with expert judgment on test scenarios, then adjust model parameters to bring the behavior of the model more in line with the experts intuition. This paper presents a methodology for analytic computation of sensitivity values in Bayesian network models. Sensitivity values are partial derivatives of output probabilities with respect to parameters being varied in the sensitivity analysis. They measure the impact of small changes in a network parameter on a target probability value or distribution. Sensitivity values can be used to focus knowledge elicitation effort on those parameters having the most impact on outputs of concern. Analytic sensitivity values are computed for an example and compared to sensitivity analysis by direct variation of parameters.< ></abstract><cop>New York, NY</cop><pub>IEEE</pub><doi>10.1109/21.384252</doi><tpages>9</tpages></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | ISSN: 0018-9472 |
| ispartof | IEEE transactions on systems, man, and cybernetics, 1995-06, Vol.25 (6), p.901-909 |
| issn | 0018-9472 2168-2909 |
| language | eng |
| recordid | cdi_pascalfrancis_primary_3533476 |
| source | IEEE Xplore |
| subjects | Applied sciences Artificial intelligence Bayesian methods Computer science control theory systems Exact sciences and technology Expert systems Intelligent networks Knowledge engineering Learning and adaptive systems Network topology Random variables Sensitivity analysis System testing Uncertainty |
| title | Sensitivity analysis for probability assessments in Bayesian networks |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-17T17%3A41%3A10IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-pascalfrancis_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Sensitivity%20analysis%20for%20probability%20assessments%20in%20Bayesian%20networks&rft.jtitle=IEEE%20transactions%20on%20systems,%20man,%20and%20cybernetics&rft.au=Laskey,%20K.B.&rft.date=1995-06-01&rft.volume=25&rft.issue=6&rft.spage=901&rft.epage=909&rft.pages=901-909&rft.issn=0018-9472&rft.eissn=2168-2909&rft.coden=ISYMAW&rft_id=info:doi/10.1109/21.384252&rft_dat=%3Cpascalfrancis_RIE%3E3533476%3C/pascalfrancis_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=384252&rfr_iscdi=true |