计算机应用 ›› 2013, Vol. 33 ›› Issue (06): 1579-1586.DOI: 10.3724/SP.J.1087.2013.01579

• 人工智能 • 上一篇    下一篇

改进的本体概念相似度计算模型

姚佳岷1,杨思春1,2   

  1. 1. 安徽工业大学 计算机学院,安徽 马鞍山 243002
    2. 南京大学 计算机软件新技术国家重点实验室,南京 210093
  • 收稿日期:2012-12-17 修回日期:2013-01-16 出版日期:2013-06-01 发布日期:2013-06-05
  • 通讯作者: 姚佳岷
  • 作者简介:姚佳岷(1986-),男,广东梅州人,硕士研究生,CCF会员,主要研究方向:概念格;杨思春(1970-),男,安徽六安人,副教授, 博士研究生,主要研究方向:自然语言处理、概念格。
  • 基金资助:

    计算机软件新技术国家重点实验室开放课题(KFKT2010B02);安徽省高校省级自然科学研究重点项目(KJ2011A48)

Improved model for similarity computation of ontology concept

YAO Jiamin1,YANG Sichun1,2   

  1. 1. School of Computer, Anhui University of Technology, Maanshan Anhui 243002, China
    2. State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing Jiangsu 210093, China
  • Received:2012-12-17 Revised:2013-01-16 Online:2013-06-05 Published:2013-06-01
  • Contact: YAO Jiamin

摘要: 本体映射能很好地解决语义网中的本体异构性问题,其核心在于计算本体概念的相似度。针对现有的概念相似度计算的精度和查准率不高,提出一种改进的概念相似度计算模型。首先利用本体特征之间的偏序关系建立形式背景和概念格,然后在结构层次求出概念间的交不可约元集,并通过对集合里各元素的语义关系进行量化计算出概念间的相似度。实例和分析结果表明,改进的概念相似度计算模型在F-Score上有明显提高。

关键词: 语义网, 本体映射, 概念相似度, 概念格

Abstract: Ontology alignment is a good solution to the problem of ontology heterogeneity in semantic Web, and its core is to calculate the similarity of ontology concepts. Given that the accuracy and precision of the existing similarity computation methods are not high, an improved model for similarity computation of ontology concept was proposed. First, this paper established a formal context and its corresponding concept lattice by using the partial order relation between ontology features, then obtained the meet-irreducible elements between concepts in structural level, and calculated similarity of concepts through quantified semantic relationships of the elements in the set. The case study and analytical results show that the improved model gets higher F-Score than other methods.

Key words: semantic Web, ontology alignment, concept similarity, concept lattice

中图分类号: