{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00048978","sets":["1164:4179:4257:4261"]},"path":["4261"],"owner":"1","recid":"48978","title":["テレビニュース番組電子化原稿を題材とした自動要約手法の大規模評価"],"pubdate":{"attribute_name":"公開日","attribute_value":"1997-05-26"},"_buckets":{"deposit":"dfce0ccb-70d8-4335-80fd-5e7af521daa3"},"_deposit":{"id":"48978","pid":{"type":"depid","value":"48978","revision_id":0},"owners":[1],"status":"published","created_by":1},"item_title":"テレビニュース番組電子化原稿を題材とした自動要約手法の大規模評価","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"テレビニュース番組電子化原稿を題材とした自動要約手法の大規模評価"},{"subitem_title":"Evaluation of Methods of Sentence Extraction on TV News Texts","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"1997-05-26","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"渋谷・上原リサーチセンター"},{"subitem_text_value":"TAO"},{"subitem_text_value":"TAO"},{"subitem_text_value":"TAO"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"TAO of Japan","subitem_text_language":"en"},{"subitem_text_value":"NHK/TAO","subitem_text_language":"en"},{"subitem_text_value":"NEC/TAO","subitem_text_language":"en"},{"subitem_text_value":"Waseda University/TAO","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/48978/files/IPSJ-NL97119006.pdf"},"date":[{"dateType":"Available","dateValue":"1999-05-26"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-NL97119006.pdf","filesize":[{"value":"344.6 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":"23"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"22fae555-5b21-486d-87bb-b33adc46d9a1","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 1997 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"若尾, 孝博"},{"creatorName":"江原暉将"},{"creatorName":"村木, 一至"},{"creatorName":"白井, 克彦"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Takahiro, Wakao","creatorNameLang":"en"},{"creatorName":"Terumasa, Ehara","creatorNameLang":"en"},{"creatorName":"Kazunori, Muraki","creatorNameLang":"en"},{"creatorName":"Katsuhiko, Shirai","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10115061","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_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"電子化された情報が増大し、テキストの要約や重要文の抽出に関する研究が、最近注目されるようになって来ている。本研究では、テレビニュース番組の電子化原稿を題材として、テキストから重要文を選び出す基本的な手法の評価を行った。基本的手法としては、重要文抽出の伝統的手法である重要語密度法、及び情報検索分野で知られているTF・IDF法に基づいた重要文抽出法を用いた。テストデータは1万件のテレビニュース番組電子化原稿で、第1文が最も重要であるというニュース原稿の特徴を利用して自動評価を行った。評価の結果は、全体的に重要度密度法がTF・IDF法に基づいた手法よりも良い結果を示すことが判明した。","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"We describe and evaluate methods for automatically selecting important sentences in TV news texts in Japanese. High-frequency word ratio method and a method based on TF-IDF are used to rank the sentences in a text. One of the features of TV news texts is that the first sentence is the most important. We take advantage of the feature and evaluate the methods automatically by using 10000 texts. On the whole, the high-frequency word ratio is better than the TF-IDF based method.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"36","bibliographic_titles":[{"bibliographic_title":"情報処理学会研究報告自然言語処理（NL）"}],"bibliographicPageStart":"31","bibliographicIssueDates":{"bibliographicIssueDate":"1997-05-26","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"53(1997-NL-119)","bibliographicVolumeNumber":"1997"}]},"relation_version_is_last":true,"weko_creator_id":"1"},"id":48978,"updated":"2025-01-22T08:14:37.527558+00:00","links":{},"created":"2025-01-18T23:14:05.855954+00:00"}