基于概念图匹配的语义相似性算法研究

本文针对传统搜索技术查全率和查准率不能满足用户日益增长的需求这一突出问题,提出一种基于概念图语义匹配的方法来计算两个本体中类之间的相似性,文中提到的本体是由实体类、这些类之间的语义关系和描述这些类的不同特征组成的.该模型首先将用户的查询信息转变为一个概念图,然后和已有的资源概念图进行匹配计算语义的相似性,实例表明该方法可以满足用户的需求,提高了检索效率.

Study on Semantic Similarity by Matching Concept Graphs

In order to solve the prominent issue that neither recall nor precision of the traditional search technology based on keyword matching can satisfy the increasing demands of users, an approach is introduced to computing the similarity of classes between two ontologies by matching concept graph, the ontology mentioned in the paper include entity class, semantic relations among these classes, and distinguish features that describe these classes. The model interpret the enquiry sentence of users into a query concept graph, then computes the similarity between the query concept graph and the resource graph, experimental results show that the technique can help user accomplish the task effectively.