基于聚类的本体块匹配方法

Clustering-Based Ontology Block Matching Approach

提出一种新的处理n∶m映射的方法,该方法将n∶m映射问题转化为聚类问题,利用Hownet中的资源使本体中的实体基于语义关系聚合,并重新给出了查全率和查准率的计算公式.使用Hownet及其相关工具对OAEI组织给出的一组本体对进行实验,实验结果表明,该方法对块匹配问题效果较好.

A new approach to deal with n ∶ m mappings was proposed,which translates the ontology matching problem into a clustering issue.It uses the information in Hownet to make the entities in ontologies cluster based on semantic relationship.The formulae of precision and recall were redefined.We used Hownet and its related tools to carry out experiments on a pair of ontologies provided by OAEI,and the experimental results demonstrate that our approach is feasible on block matching problem.