A Wikipedia based hybrid ranking method for taxonomic relation extraction

Xiaoshi Zhong*

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

This paper proposes a hybrid ranking method for taxonomic relation extraction (or select best position) in an existing taxonomy. This method is capable of effectively combining two resources, an existing taxonomy and Wikipedia, in order to select a most appropriate position for a term candidate in the existing taxonomy. Previous methods mainly focus on complex inference methods to select the best position among all the possible position in the taxonomy. In contrast, our algorithm, a simple but effective one, leverage two kinds of information, the expression of and the ranking information of a term candidate, to select the best position for the term candidate (the hypernym of the term candidate in the existing taxonomy). We conduct our approach on the agricultural domain and the experimental result indicates that the performances are significantly improved.

源语言英语
主期刊名Information Retrieval Technology - 9th Asia Information Retrieval Societies Conference, AIRS 2013, Proceedings
332-343
页数12
DOI
出版状态已出版 - 2013
已对外发布
活动9th Asia Information Retrieval Societies Conference on Information Retrieval Technology, AIRS 2013 - Singapore, 新加坡
期限: 9 12月 201311 12月 2013

出版系列

姓名Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
8281 LNCS
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议9th Asia Information Retrieval Societies Conference on Information Retrieval Technology, AIRS 2013
国家/地区新加坡
Singapore
时期9/12/1311/12/13

指纹

探究 'A Wikipedia based hybrid ranking method for taxonomic relation extraction' 的科研主题。它们共同构成独一无二的指纹。

引用此