基于大规模本体的有效匹配研究

Effective Matching Based on Large Scale Ontology

为了改善不同领域专家构建的本体存在不匹配的问题,以及大规模本体问题已经超过当前的匹配方法,笔者设计了一种基于模块化的方法,其思路是把一个大的匹配问题分解成几个小的匹配问题,以减少急剧变化的复杂性.为了验证该方法,选定了几个较大和复杂的本体来测试.结果表明,基于模块化的方法(MOM)在保持较好的准确率和查全率的同时还可以较大地减少匹配时间.

In order to solve the problems of the mismatching of ontology constructed by the .experts in different fields and that the large scale ontology problems is beyond the reach of the existing matching methods, a modularization-based approach was designed. The main idea of the design was to decompose a large matching problem into several smaller ones and reduce the complexity of the dramatically change. To further verify the approach, several large and complex ontology had been chosen to be tested. The results showed that this method could significantly reduce the time cost and keep the high matching accuracy and recall.