{"created":"2025-01-19T01:19:22.588809+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00219042","sets":["1164:1384:10896:10984"]},"path":["10984"],"owner":"44499","recid":"219042","title":["連合学習によるプライバシー保護を考慮したプロジェクト間バグ予測"],"pubdate":{"attribute_name":"公開日","attribute_value":"2022-07-21"},"_buckets":{"deposit":"8ec76000-24af-4e03-aa32-89ba7d748628"},"_deposit":{"id":"219042","pid":{"type":"depid","value":"219042","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"連合学習によるプライバシー保護を考慮したプロジェクト間バグ予測","author_link":["570870","570873","570872","570871","570874"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"連合学習によるプライバシー保護を考慮したプロジェクト間バグ予測"}]},"item_type_id":"4","publish_date":"2022-07-21","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"九州大学"},{"subitem_text_value":"Centre for Software Excellence"},{"subitem_text_value":"九州大学"},{"subitem_text_value":"九州大学"},{"subitem_text_value":"九州大学"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Kyushu University","subitem_text_language":"en"},{"subitem_text_value":"Centre for Software Excellence","subitem_text_language":"en"},{"subitem_text_value":"Kyushu University","subitem_text_language":"en"},{"subitem_text_value":"Kyushu University","subitem_text_language":"en"},{"subitem_text_value":"Kyushu University","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"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/219042/files/IPSJ-SE22211016.pdf","label":"IPSJ-SE22211016.pdf"},"date":[{"dateType":"Available","dateValue":"2024-07-21"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SE22211016.pdf","filesize":[{"value":"1.1 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"12"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"b6efba58-fc57-4ce2-a985-3390c314d521","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2022 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"山本, 大貴"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Gopi, Krishnan Rajbahadur"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"近藤, 将成"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"亀井, 靖高"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"鵜林, 尚靖"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10112981","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8825","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"バグ予測モデルはソフトウェアプロジェクトに含まれるバグの発生率を予測できるため，デバッグ作業の支援の手法として活用されている．近年，バグ予測モデルを作成するためのデータが無い場合に，他人の持つデータを利用して学習したバグ予測モデルを用いてプロジェクトのバグを予測する Cross Project Defect Prediction（CPDP）への関心が高まっている．CPDP では他人のデータを利用することから，データのプライバシー保護が重要な課題の 1 つである．しかし，多くの CPDP の研究ではモデルの学習のためにプロジェクト間でのデータの共有が必要であり，プロジェクトの機密性を保護しながらプロジェクト間のバグを予測することを考慮していない．そこで本研究では，データ共有が不要な分散型の機械学習アプローチである連合学習（Federated Learning）を用いた CPDP モデルを提案し，CPDP に連合学習のモデルを利用した際の性能や特徴を明らかにする．連合学習の CPDP モデルと既存の CPDP モデルとの性能比較を AUC を用いて行った結果として，連合学習のモデルは実験に用いたデータセットの 68% で予測性能の順位グループの 2 位以上に割り当てられ，既存のモデルと比較しても一定の予測性能があることを明らかにした．また，本研究の実験環境における連合学習のモデルの予測性能に及ぼす影響が大きい特徴量を調査し，本質的複雑度やコード行数のようなコードに関する特徴量は性能に影響を及ぼすことを明らかにした．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"8","bibliographic_titles":[{"bibliographic_title":"研究報告ソフトウェア工学（SE）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2022-07-21","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"16","bibliographicVolumeNumber":"2022-SE-211"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":219042,"updated":"2025-01-19T14:56:44.942785+00:00","links":{}}