A code change‐oriented approach to just‐in‐time defect prediction with multiple input semantic fusion

Recent research found that fine‐tuning pre‐trained models is superior to training models from scratch in just‐in‐time (JIT) defect prediction. However, existing approaches using pre‐trained models have their limitations. First, the input length is constrained by the pre‐trained models.Secondly, the...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Expert systems 2024-12, Vol.41 (12), p.n/a
Hauptverfasser: Huang, Teng, Yu, Hui‐Qun, Fan, Gui‐Sheng, Huang, Zi‐Jie, Wu, Chen‐Yu
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page n/a
container_issue 12
container_start_page
container_title Expert systems
container_volume 41
creator Huang, Teng
Yu, Hui‐Qun
Fan, Gui‐Sheng
Huang, Zi‐Jie
Wu, Chen‐Yu
description Recent research found that fine‐tuning pre‐trained models is superior to training models from scratch in just‐in‐time (JIT) defect prediction. However, existing approaches using pre‐trained models have their limitations. First, the input length is constrained by the pre‐trained models.Secondly, the inputs are change‐agnostic.To address these limitations, we propose JIT‐Block, a JIT defect prediction method that combines multiple input semantics using changed block as the fundamental unit. We restructure the JIT‐Defects4J dataset used in previous research. We then conducted a comprehensive comparison using eleven performance metrics, including both effort‐aware and effort‐agnostic measures, against six state‐of‐the‐art baseline models. The results demonstrate that on the JIT defect prediction task, our approach outperforms the baseline models in all six metrics, showing improvements ranging from 1.5% to 800% in effort‐agnostic metrics and 0.3% to 57% in effort‐aware metrics. For the JIT defect code line localization task, our approach outperforms the baseline models in three out of five metrics, showing improvements of 11% to 140%.
doi_str_mv 10.1111/exsy.13702
format Article
fullrecord <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_3132449285</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3132449285</sourcerecordid><originalsourceid>FETCH-LOGICAL-c2262-b39a8711579f397bf53bb209e14a991807bd80f0bd0b6617d46088fb30ff00583</originalsourceid><addsrcrecordid>eNp9kL1OwzAQgC0EEqWw8ASW2JBSznYSO2NVlR-pEgMgwWQ5iU1d8kfsqHTjEXhGngSXMHPD3XDf3ek-hM4JzEiIK_3hdjPCONADNCFxKiJgWXyIJkDTNIo5hWN04twGAAjn6QS9zXHRlhoXa9W86u_Pr7a3uvG6xKrr-lYVa-xbvBmcDz3bhORtrXGpjS487npd2sLbtsFb69e4Hipvu0pj23SDx07XqvG2wGZwgTlFR0ZVTp_91Sl6ul4-Lm6j1f3N3WK-igpKUxrlLFOCE5LwzLCM5yZheU4h0yRWWUYE8LwUYCAvIU9Twss4BSFMzsAYgESwKboY94YH3gftvNy0Q9-Ek5IRRuM4oyIJ1OVIFX3rXK-N7Hpbq34nCci9TLmXKX9lBpiM8NZWevcPKZfPDy_jzA8bs3rx</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3132449285</pqid></control><display><type>article</type><title>A code change‐oriented approach to just‐in‐time defect prediction with multiple input semantic fusion</title><source>Access via Wiley Online Library</source><creator>Huang, Teng ; Yu, Hui‐Qun ; Fan, Gui‐Sheng ; Huang, Zi‐Jie ; Wu, Chen‐Yu</creator><creatorcontrib>Huang, Teng ; Yu, Hui‐Qun ; Fan, Gui‐Sheng ; Huang, Zi‐Jie ; Wu, Chen‐Yu</creatorcontrib><description>Recent research found that fine‐tuning pre‐trained models is superior to training models from scratch in just‐in‐time (JIT) defect prediction. However, existing approaches using pre‐trained models have their limitations. First, the input length is constrained by the pre‐trained models.Secondly, the inputs are change‐agnostic.To address these limitations, we propose JIT‐Block, a JIT defect prediction method that combines multiple input semantics using changed block as the fundamental unit. We restructure the JIT‐Defects4J dataset used in previous research. We then conducted a comprehensive comparison using eleven performance metrics, including both effort‐aware and effort‐agnostic measures, against six state‐of‐the‐art baseline models. The results demonstrate that on the JIT defect prediction task, our approach outperforms the baseline models in all six metrics, showing improvements ranging from 1.5% to 800% in effort‐agnostic metrics and 0.3% to 57% in effort‐aware metrics. For the JIT defect code line localization task, our approach outperforms the baseline models in three out of five metrics, showing improvements of 11% to 140%.</description><identifier>ISSN: 0266-4720</identifier><identifier>EISSN: 1468-0394</identifier><identifier>DOI: 10.1111/exsy.13702</identifier><language>eng</language><publisher>Oxford: Blackwell Publishing Ltd</publisher><subject>deep learning ; defect prediction ; Defects ; just‐in‐time ; Performance measurement ; Semantics ; software defect</subject><ispartof>Expert systems, 2024-12, Vol.41 (12), p.n/a</ispartof><rights>2024 John Wiley &amp; Sons Ltd.</rights><rights>2024 John Wiley &amp; Sons, Ltd.</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c2262-b39a8711579f397bf53bb209e14a991807bd80f0bd0b6617d46088fb30ff00583</cites><orcidid>0009-0002-3909-6778</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://onlinelibrary.wiley.com/doi/pdf/10.1111%2Fexsy.13702$$EPDF$$P50$$Gwiley$$H</linktopdf><linktohtml>$$Uhttps://onlinelibrary.wiley.com/doi/full/10.1111%2Fexsy.13702$$EHTML$$P50$$Gwiley$$H</linktohtml><link.rule.ids>315,781,785,1418,27929,27930,45579,45580</link.rule.ids></links><search><creatorcontrib>Huang, Teng</creatorcontrib><creatorcontrib>Yu, Hui‐Qun</creatorcontrib><creatorcontrib>Fan, Gui‐Sheng</creatorcontrib><creatorcontrib>Huang, Zi‐Jie</creatorcontrib><creatorcontrib>Wu, Chen‐Yu</creatorcontrib><title>A code change‐oriented approach to just‐in‐time defect prediction with multiple input semantic fusion</title><title>Expert systems</title><description>Recent research found that fine‐tuning pre‐trained models is superior to training models from scratch in just‐in‐time (JIT) defect prediction. However, existing approaches using pre‐trained models have their limitations. First, the input length is constrained by the pre‐trained models.Secondly, the inputs are change‐agnostic.To address these limitations, we propose JIT‐Block, a JIT defect prediction method that combines multiple input semantics using changed block as the fundamental unit. We restructure the JIT‐Defects4J dataset used in previous research. We then conducted a comprehensive comparison using eleven performance metrics, including both effort‐aware and effort‐agnostic measures, against six state‐of‐the‐art baseline models. The results demonstrate that on the JIT defect prediction task, our approach outperforms the baseline models in all six metrics, showing improvements ranging from 1.5% to 800% in effort‐agnostic metrics and 0.3% to 57% in effort‐aware metrics. For the JIT defect code line localization task, our approach outperforms the baseline models in three out of five metrics, showing improvements of 11% to 140%.</description><subject>deep learning</subject><subject>defect prediction</subject><subject>Defects</subject><subject>just‐in‐time</subject><subject>Performance measurement</subject><subject>Semantics</subject><subject>software defect</subject><issn>0266-4720</issn><issn>1468-0394</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNp9kL1OwzAQgC0EEqWw8ASW2JBSznYSO2NVlR-pEgMgwWQ5iU1d8kfsqHTjEXhGngSXMHPD3XDf3ek-hM4JzEiIK_3hdjPCONADNCFxKiJgWXyIJkDTNIo5hWN04twGAAjn6QS9zXHRlhoXa9W86u_Pr7a3uvG6xKrr-lYVa-xbvBmcDz3bhORtrXGpjS487npd2sLbtsFb69e4Hipvu0pj23SDx07XqvG2wGZwgTlFR0ZVTp_91Sl6ul4-Lm6j1f3N3WK-igpKUxrlLFOCE5LwzLCM5yZheU4h0yRWWUYE8LwUYCAvIU9Twss4BSFMzsAYgESwKboY94YH3gftvNy0Q9-Ek5IRRuM4oyIJ1OVIFX3rXK-N7Hpbq34nCci9TLmXKX9lBpiM8NZWevcPKZfPDy_jzA8bs3rx</recordid><startdate>202412</startdate><enddate>202412</enddate><creator>Huang, Teng</creator><creator>Yu, Hui‐Qun</creator><creator>Fan, Gui‐Sheng</creator><creator>Huang, Zi‐Jie</creator><creator>Wu, Chen‐Yu</creator><general>Blackwell Publishing Ltd</general><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>7TB</scope><scope>8FD</scope><scope>F28</scope><scope>FR3</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><orcidid>https://orcid.org/0009-0002-3909-6778</orcidid></search><sort><creationdate>202412</creationdate><title>A code change‐oriented approach to just‐in‐time defect prediction with multiple input semantic fusion</title><author>Huang, Teng ; Yu, Hui‐Qun ; Fan, Gui‐Sheng ; Huang, Zi‐Jie ; Wu, Chen‐Yu</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c2262-b39a8711579f397bf53bb209e14a991807bd80f0bd0b6617d46088fb30ff00583</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>deep learning</topic><topic>defect prediction</topic><topic>Defects</topic><topic>just‐in‐time</topic><topic>Performance measurement</topic><topic>Semantics</topic><topic>software defect</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Huang, Teng</creatorcontrib><creatorcontrib>Yu, Hui‐Qun</creatorcontrib><creatorcontrib>Fan, Gui‐Sheng</creatorcontrib><creatorcontrib>Huang, Zi‐Jie</creatorcontrib><creatorcontrib>Wu, Chen‐Yu</creatorcontrib><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Mechanical &amp; Transportation Engineering Abstracts</collection><collection>Technology Research Database</collection><collection>ANTE: Abstracts in New Technology &amp; Engineering</collection><collection>Engineering Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>Expert systems</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Huang, Teng</au><au>Yu, Hui‐Qun</au><au>Fan, Gui‐Sheng</au><au>Huang, Zi‐Jie</au><au>Wu, Chen‐Yu</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>A code change‐oriented approach to just‐in‐time defect prediction with multiple input semantic fusion</atitle><jtitle>Expert systems</jtitle><date>2024-12</date><risdate>2024</risdate><volume>41</volume><issue>12</issue><epage>n/a</epage><issn>0266-4720</issn><eissn>1468-0394</eissn><abstract>Recent research found that fine‐tuning pre‐trained models is superior to training models from scratch in just‐in‐time (JIT) defect prediction. However, existing approaches using pre‐trained models have their limitations. First, the input length is constrained by the pre‐trained models.Secondly, the inputs are change‐agnostic.To address these limitations, we propose JIT‐Block, a JIT defect prediction method that combines multiple input semantics using changed block as the fundamental unit. We restructure the JIT‐Defects4J dataset used in previous research. We then conducted a comprehensive comparison using eleven performance metrics, including both effort‐aware and effort‐agnostic measures, against six state‐of‐the‐art baseline models. The results demonstrate that on the JIT defect prediction task, our approach outperforms the baseline models in all six metrics, showing improvements ranging from 1.5% to 800% in effort‐agnostic metrics and 0.3% to 57% in effort‐aware metrics. For the JIT defect code line localization task, our approach outperforms the baseline models in three out of five metrics, showing improvements of 11% to 140%.</abstract><cop>Oxford</cop><pub>Blackwell Publishing Ltd</pub><doi>10.1111/exsy.13702</doi><tpages>16</tpages><orcidid>https://orcid.org/0009-0002-3909-6778</orcidid><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier ISSN: 0266-4720
ispartof Expert systems, 2024-12, Vol.41 (12), p.n/a
issn 0266-4720
1468-0394
language eng
recordid cdi_proquest_journals_3132449285
source Access via Wiley Online Library
subjects deep learning
defect prediction
Defects
just‐in‐time
Performance measurement
Semantics
software defect
title A code change‐oriented approach to just‐in‐time defect prediction with multiple input semantic fusion
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-10T00%3A17%3A32IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=A%20code%20change%E2%80%90oriented%20approach%20to%20just%E2%80%90in%E2%80%90time%20defect%20prediction%20with%20multiple%20input%20semantic%20fusion&rft.jtitle=Expert%20systems&rft.au=Huang,%20Teng&rft.date=2024-12&rft.volume=41&rft.issue=12&rft.epage=n/a&rft.issn=0266-4720&rft.eissn=1468-0394&rft_id=info:doi/10.1111/exsy.13702&rft_dat=%3Cproquest_cross%3E3132449285%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3132449285&rft_id=info:pmid/&rfr_iscdi=true