{"id":231555,"updated":"2025-01-19T10:43:08.743448+00:00","links":{},"created":"2025-01-19T01:31:55.584838+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00231555","sets":["581:11107:11121"]},"path":["11121"],"owner":"44499","recid":"231555","title":["事前学習モデルを利用したソースコード片の不具合予測"],"pubdate":{"attribute_name":"公開日","attribute_value":"2023-12-15"},"_buckets":{"deposit":"e425bcd0-c6a6-4a07-a310-aca547afb5e1"},"_deposit":{"id":"231555","pid":{"type":"depid","value":"231555","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"事前学習モデルを利用したソースコード片の不具合予測","author_link":["625259","625258","625257","625256"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"事前学習モデルを利用したソースコード片の不具合予測"},{"subitem_title":"Defect Prediction Based on Source Code Fragments Using Pre-trained Model","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"[一般論文] 不具合予測，深層学習，事前学習","subitem_subject_scheme":"Other"}]},"item_type_id":"2","publish_date":"2023-12-15","item_2_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"龍谷大学大学院理工学研究科"},{"subitem_text_value":"龍谷大学先端理工学部"}]},"item_2_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Science and Technology, Rykoku University","subitem_text_language":"en"},{"subitem_text_value":"Faculty of Advance Science and Technology, Rykoku University","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/231555/files/IPSJ-JNL6412013.pdf","label":"IPSJ-JNL6412013.pdf"},"date":[{"dateType":"Available","dateValue":"2025-12-15"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-JNL6412013.pdf","filesize":[{"value":"760.0 kB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"8"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"5de714a3-d336-4713-afd1-303cd694db78","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2023 by the Information Processing Society of Japan"}]},"item_2_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"河方, 健悟"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"山本, 哲男"}],"nameIdentifiers":[{}]}]},"item_2_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Kengo, Kawakata","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Tetsuo, Yamamoto","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_2_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN00116647","subitem_source_identifier_type":"NCID"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_6501","resourcetype":"journal article"}]},"item_2_publisher_15":{"attribute_name":"公開者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"item_2_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"1882-7764","subitem_source_identifier_type":"ISSN"}]},"item_2_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"ソフトウェアの品質を確保するために不具合の含まれる可能性の高いソースコードを予測し，予測された箇所を重点的にテストすることは，保守活動の信頼性向上，効率化のために重要である．そのための手法として，深層学習を用いてソースコード内に含まれる不具合を予測する方法がある．本研究では，実験のためGitリポジトリのオープンソースソフトウェアからソースコード片を収集し，それらに対して深層学習を用いて不具合が含まれるか否かを予測する．深層学習のモデルには，先行研究によってソースファイル単位での不具合予測において有効であると示された，ソースコードの事前学習を行ったモデルであるCodeBERTと自然言語のみのモデルであるRoBERTaを用いて，ソースコード片に対して不具合予測を行い評価した．評価実験によって，事前学習ありのモデルはソースコード片の不具合予測において有効であること，事前学習なしのモデルに比べて精度が高いこと，さらに，事前学習に使用されたプログラミング言語のソースコード片においても有効であることを示す．","subitem_description_type":"Other"}]},"item_2_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"It is important to predict source code defects and to test the predicted areas intensively in order to improve the reliability and efficiency of maintenance activities. There are methods using deep learning for this purpose to predict defects in source code. In this study, we collect source code fragments from open source software in Git repositories for experiments, and use deep learning to predict whether or not they contain defects. We used CodeBERT, a model with pre-training on source code, and RoBERTa, a model with pre-training on only natural language, both of which have been shown to be effective in predicting defects per source file in previous studies. Through evaluation experiments, we show that the model with pre-training is effective in predicting defects in source code fragments, is more accurate than the model without pre-training, and is also effective in source code fragments in the programming language used for pre-training.","subitem_description_type":"Other"}]},"item_2_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"1667","bibliographic_titles":[{"bibliographic_title":"情報処理学会論文誌"}],"bibliographicPageStart":"1659","bibliographicIssueDates":{"bibliographicIssueDate":"2023-12-15","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"12","bibliographicVolumeNumber":"64"}]},"relation_version_is_last":true,"item_2_identifier_registration":{"attribute_name":"ID登録","attribute_value_mlt":[{"subitem_identifier_reg_text":"10.20729/00231445","subitem_identifier_reg_type":"JaLC"}]},"weko_creator_id":"44499"}}