基于实例相似度的概念语义挖掘方法

摘要

Ontology plays an important role in knowledge expression, sharing, reuse and semantic query, but in the process of ontology integration,it is difficult to fuse the concept layer. Aiming at this paper,this paper proposes a method to find concept semantic for ontology. Based on the instance similarity metric extract from instance data, it alternates between instance level and schema level to get concept semantic, mines the property semantic, and then turns instance level to amend similarity metric and complement concept semantic on schema level. This paper tests this method with test ontologies provided by OAEI2012,and the result shows that the precise and recall ratio are increased,which proves the feasibility and effectiveness of the method.%本体在知识表达、共享、重用以及语义查询中具有重要作用，但在本体融合过程中存在概念层融合难的问题。为此，提出一种挖掘本体概念语义的方法。该方法从实例数据出发，以实例相似度矩阵为基础，在实例层和概念层交替进行概念语义挖掘，将挖掘结果通过属性语义反馈到实例层，并对其进行修正和补充。利用OAEI2012提供的测试本体进行实验，结果显示查全率与查准率均得到提高，证明了该方法的可行性和有效性。