一种基于混合语义相似度度量方法的本体元匹配技术

An ontology meta-matching technique based on the hybrid semantic similarity measure

目的通过混合多种语义相似度度量方法来提高本体匹配的质量。方法构建一种新的语义相似度度量方法的组合框架,综合考虑组合过程中存在的2种情况来提高混合语义相似度度量方法的有效性。结果通过同国际本体匹配结果评价竞赛(Ontology Alignment Evaluation Initiative,OAEI)参与者的比较结果表明本文所提出的方法是有效的。结论本文提出的混合语义相似度度量方法能够有效组合基于知识库和字符串的语义相似度度量方法以识别异质的本体概念,从而提高本体匹配结果的质量。

Purposes—To improve the quality of the ontology alignment by integrating various semantic similarity measures.Methods—A new combination framework of semantic similarity measure is constructed,which ensure the effectiveness of the hybrid semantic similarity measure by taking into two distinct situations consideration.Result—The comparisons with the participants of Ontology Alignment Evaluation Initiative(OAEI)show the effectiveness of our proposal.Conclusion—The proposed hybrid semantic similarity measure can effectively combine knowledge-based and string-based semantic similarity measures to identifying the heterogeneous ontology concepts,thus,improving the quality of ontology alignments.