摘要
针对现有本体映射中基于元素名称匹配的方法有效性不高等缺点,设计一种综合的元素名称特征匹配的本体映射方法.该方法引入了词义消歧算法,以确定元素名称词汇在本体上下文中表示的确切含义,并设计了更合理的基于信息量和基于语义边加权的概念语义相似度计算方法.通过选择最佳的词形相似度方法,并适当地组合语义相似度和词形相似度,实现了更有效的本体元素名称匹配策略.实验结果表明,本文方法能够在保持查准率的基础上,明显提高本体映射的查全率.
In view of the low effectiveness of elements name matching methods of existing ontology mapping systems,a synthetical elements name feature matching based approach for ontology mapping is proposed. In this approach,w e introduced the w ord sense disambiguation algorithm to analyze the elements name's real meaning,and a more reasonable concept semantic similarity metric w as designed,w hich depends on w eighting information content and semantic edge based method. Through chose morphological similarity metric that has the best performance,and combined it w ith the new semantic similarity metric,w e achieved more effective ontology elements name matching. The experimental results show that,this approach can improve the ontology mapping's recall obviously on the basis of keeping high precision.
出处
《小型微型计算机系统》
CSCD
北大核心
2015年第9期2009-2014,共6页
Journal of Chinese Computer Systems
基金
总装预研基金项目(9140A27040413JB11407)资助
国家自然科学基金项目(61170217)资助
关键词
本体映射
名称匹配
词义消歧
语义相似度
ontology mapping
name matching
w ord sense disambiguation
semantic similarity