{"created":"2025-01-19T01:09:55.958944+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00208546","sets":["6164:6165:6462:10428"]},"path":["10428"],"owner":"44499","recid":"208546","title":["セキュリティレポートのマルチラベル分類のためのトピックモデルの汎化性能に着目した外れ値検出の適用"],"pubdate":{"attribute_name":"公開日","attribute_value":"2020-10-19"},"_buckets":{"deposit":"1cd6289c-fac2-423e-8273-20238c62b19b"},"_deposit":{"id":"208546","pid":{"type":"depid","value":"208546","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"セキュリティレポートのマルチラベル分類のためのトピックモデルの汎化性能に着目した外れ値検出の適用","author_link":["523540","523538","523535","523547","523542","523537","523545","523548","523549","523539","523541","523534","523546","523544","523543","523536"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"セキュリティレポートのマルチラベル分類のためのトピックモデルの汎化性能に着目した外れ値検出の適用"},{"subitem_title":"Applying Outlier Detection to Improving Modeling Accuracy of Topic Models for Multi-Labeling Security Reports","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"脅威情報，マルチラベリング，トピックモデル，外れ値検出，セキュリティレポート","subitem_subject_scheme":"Other"}]},"item_type_id":"18","publish_date":"2020-10-19","item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_18_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"神戸大学 大学院工学研究科"},{"subitem_text_value":"兵庫県立大学 社会情報科学部"},{"subitem_text_value":"国立研究開発法人 情報通信研究機構"},{"subitem_text_value":"神戸大学 大学院工学研究科"},{"subitem_text_value":"国立研究開発法人 情報通信研究機構"},{"subitem_text_value":"岐阜大学 工学部電気電子・情報工学科"},{"subitem_text_value":"神戸大学 大学院工学研究科"},{"subitem_text_value":"神戸大学 大学院工学研究科"}]},"item_18_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Kobe University","subitem_text_language":"en"},{"subitem_text_value":"University of Hyogo","subitem_text_language":"en"},{"subitem_text_value":"National Institute of Information and Communications Technology","subitem_text_language":"en"},{"subitem_text_value":"Kobe University","subitem_text_language":"en"},{"subitem_text_value":"National Institute of Information and Communications Technology","subitem_text_language":"en"},{"subitem_text_value":"Gifu University","subitem_text_language":"en"},{"subitem_text_value":"Kobe University","subitem_text_language":"en"},{"subitem_text_value":"Kobe University","subitem_text_language":"en"}]},"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/208546/files/IPSJCSS2020118.pdf","label":"IPSJCSS2020118.pdf"},"date":[{"dateType":"Available","dateValue":"2022-10-19"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJCSS2020118.pdf","filesize":[{"value":"461.3 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":"30"},{"tax":["include_tax"],"price":"0","billingrole":"46"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"e53673d2-df64-4f43-b9af-87c363ddc81a","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2020 by the Information Processing Society of Japan"}]},"item_18_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"長田, 侑樹"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"瀧田, 愼"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"古本, 啓祐"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"白石, 善明"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"高橋, 健志"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"毛利, 公美"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"髙野, 泰洋"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"森井, 昌克"}],"nameIdentifiers":[{}]}]},"item_18_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Yuki, Osada","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Makoto, Takita","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Keisuke, Furumoto","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Yoshiaki, Shiraishi","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Takeshi, Takahashi","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Masami, Mohri","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Yasuhiro, Takano","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Masakatu, Morii","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_5794","resourcetype":"conference paper"}]},"item_18_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"セキュリティレポートに付与されるラベルは発行元によって異なっており，自組織に関連するレポートを見つけ出すのは容易ではない．そのため，日々増加していくセキュリティレポートから所望の情報を得るために，レポートの分類と高精度なラベル付けが望まれている．しかし文書分類を行う際，内容が他と大きく異なる文書が多数存在する場合，クラスタリング精度が低下する可能性がある．本稿では，文書クラスタリングの精度向上を目的として，他の文書と内容の異なる文書を外れ値文書と見なし，外れ値文書を除いてトピックモデルを構築することを提案する．ケーススタディとしてセキュリティベンダー8社が2017年から2019年に発行した2386件のセキュリティレポートに対して提案手法を適用し，トピックモデルの評価値であるPerplexityを考慮することでトピックモデルの汎化性能が向上することを確認した．また，汎化性能の向上に伴い，文書クラスタリングの精度が向上することを確認した．","subitem_description_type":"Other"}]},"item_18_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"The labels given to security reports vary from one publisher to another and it is not easy to find the reports that are relevant to your organization. Therefore, in order to obtain the desired information from the security reports, there is a need to classify and label the reports with high accuracy. However, if there are many documents whose contents are very different from those of others, the clustering accuracy may decrease. In this paper, we propose to apply outlier detection to topic models. By constructing topic models excluding outlier documents, we can improve the generalization ability of topic models. As a case study, we applied the proposed method to 2,386 security reports published by eight security vendors, and confirmed that the modeling accuracy of the topic model can be improved by considering Perplexity. We also confirmed that the accuracy of document clustering improved with the improvement of modeling accuracy.","subitem_description_type":"Other"}]},"item_18_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"852","bibliographic_titles":[{"bibliographic_title":"コンピュータセキュリティシンポジウム2020論文集"}],"bibliographicPageStart":"847","bibliographicIssueDates":{"bibliographicIssueDate":"2020-10-19","bibliographicIssueDateType":"Issued"}}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":208546,"updated":"2025-01-19T18:49:04.950375+00:00","links":{}}