北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2013, Vol. 36 ›› Issue (5): 81-85.doi: 10.13190/j.jbupt.2013.05.017

• 研究报告 • 上一篇    下一篇

特定领域概念间关系自动抽取方法

杨月华1,2,3, 杜军平1,2, 平源3   

  1. 1. 北京邮电大学 计算机学院, 北京 100876;
    2. 北京邮电大学 智能通信软件与多媒体北京市重点实验室, 北京 100876;
    3. 许昌学院 信息工程学院, 河南 许昌 461000
  • 收稿日期:2012-12-24 出版日期:2013-10-31 发布日期:2013-08-09
  • 作者简介:杨月华(1983-),女,博士生,E-mail:yangyuehua0504@126.com;杜军平(1963-),女,教授,博士生导师.
  • 基金资助:

    国家自然科学基金项目(91024001,61070142,61303232);北京市自然科学基金项目(4111002);河南省教育厅基金项目(13A413750,13A413747);河南省自然科学基金项目(132300410393)

Hybrid Automatic Extraction Method of Relationships Between Specific Domain Concepts

YANG Yue-hua1,2,3, DU Jun-ping1,2, PING Yuan3   

  1. 1. School of Computer Science, Beijing University of Posts and Telecommunications, Beijing 100876, China;
    2. Beijing Key Laboratory of Intelligent Telecommunication Software and Multimedia, Beijing University of Posts and Telecommunications, Beijing 100876, China;
    3. School of Information Engineering, Xuchang University, Henan Xuchang 461000, China
  • Received:2012-12-24 Online:2013-10-31 Published:2013-08-09

摘要:

以突发事件领域为例,针对本体构建中领域概念间关系自动抽取的问题,提出了一种混合的领域概念间关系自动抽取方法,将领域概念间的关系分成关系类型未知和已知2种情况,并分别基于扩展关联规则和关系抽取规则进行抽取,同时提出了构造和自动扩展关系抽取规则的方法。实验结果表明,所提出的方法是可行和有效的,不仅能获得特定领域中存在的丰富的语义关系,而且能获得较好的性能。

关键词: 混合关系抽取算法, 扩展关联规则, 抽取规则

Abstract:

Aiming at automatic extraction of relationships between domain concepts in ontology construction, a hybrid automatic extraction method is proposed. The relationships existing in domain concepts are classified into two kinds: one is the relationships with unknown types extracted by the extended association rule; the other is the relationships with the known types extracted by extraction rules. Meanwhile, the formation and expansion methods of extraction rules are presented. Experiments demonstrate that the proposed method is feasible and effective. It can not only obtain rich relationships in special domain, but also obtain better performance.

Key words: hybrid extraction algorithm of relationships, extended association rule, extraction rule