Raising the Compatibility of Heterogeneous Annotations : A Case Study on Protein Mention Recognition

• Item type: SIG Technical Reports(1)
• 公開日: 2008-05-15
• タイトル: Raising the Compatibility of Heterogeneous Annotations : A Case Study on Protein Mention Recognition
• タイトル / 言語: en
• タイトル: Raising the Compatibility of Heterogeneous Annotations : A Case Study on Protein Mention Recognition
• 言語: eng
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: Department of Computer Science University of Tokyo
• 著者所属: Department of Computer Science University of Tokyo
• 著者所属: Department of Computer Science University of Tokyo
• 著者所属: Department of Computer Science University of Tokyo
• 著者所属: Department of Computer Science University
• 著者所属(英): en
• 著者所属(英): Department of Computer Science, University of Tokyo
• 著者所属(英): en
• 著者所属(英): Department of Computer Science, University of Tokyo
• 著者所属(英): en
• 著者所属(英): Department of Computer Science, University of Tokyo
• 著者所属(英): en
• 著者所属(英): Department of Computer Science, University of Tokyo
• 著者所属(英): en
• 著者所属(英): Department of Computer Science, University of Tokyo / School of Informatics, University of Manchester / National Center for Text Mining
• 著者名: Yue, Wang; Kazuhiro, Yoshida; Jin-DongKim; Rune, S?tre; Jun'ichi, Tsujii
• 著者名(英): Yue, Wang; Kazuhiro, Yoshida; Jin-Dong, Kim; Rune, Satre; Jun'ichi, Tsujii
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: The use of human-annotated corpora is popular in developing language processing systems. For bio-text mining for example there are several well-known corpora with protein mention annotations. Because of the different conventions adopted by these corpora one problem that is well recognized but yet less addressed is brought about; the problem is the heterogeneity of the corpora. The problem weakens the protein annotation consistency. In this work we seek a way of removing or relaxing the heterogeneity of annotations by identifying and removing the specific difference between the annotations. Our results show that our effort to remove several disagreements between the corpora annotation is successful in reducing the performance degradation caused by heterogeneity and incompatibility.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: The use of human-annotated corpora is popular in developing language processing systems. For bio-text mining, for example, there are several well-known corpora with protein mention annotations. Because of the different conventions adopted by these corpora, one problem that is well recognized, but yet less addressed is brought about; the problem is the heterogeneity of the corpora. The problem weakens the protein annotation consistency. In this work, we seek a way of removing or relaxing the heterogeneity of annotations by identifying and removing the specific difference between the annotations. Our results show that our effort to remove several disagreements between the corpora annotation is successful in reducing the performance degradation caused by heterogeneity and incompatibility.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AN10115061
• 書誌情報: 情報処理学会研究報告自然言語処理（NL）; 巻 2008, 号 46(2008-NL-185), p. 53-58, 発行日 2008-05-15
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会