摘要： 提出一种基于多属性的本体实例匹配算法，利用关键性属性获得待匹配本体实例间的相似度，使用Similarity Flooding算法，依据由对象属性构成的有向图对实例间的相似度进行修正，并获得初始匹配结果，根据必要性属性过滤初始匹配中的错误匹配。实验结果表明，该算法可提高本体实例匹配的查准率与查全率。

关键词: 本体, 本体实例匹配, 语义网, 多属性

Abstract: This paper proposes an ontology instance matching algorithm based on multi-property. The approach obtains the similarities between instances. And then for the initial matching results, the method uses the Similarity Flooding algorithm to revise these similarities according to the directed graph constructed by the object properties. The necessity properties are employed to filter out the incorrected matching pairs. Experimental results show that the proposed method can increase the precision ratio and recall ratio of ontology instance matching.

Key words: ontology, ontology instance matching, semantic Web, multi-property

中图分类号: 

• TP391