一种基于知网的概念相似度计算方法

概念相似度计算是句子相似度计算的基础,以知网中的义原关系为依据,认为两个概念,其独立义原对相似度大小起着决定性作用;独立义原越相近,描述的信息差越小,它们的相似度也就越大.提出了一种基于知网的概念相似度的计算方法,定义了概念相似度公式.实验表明,据此计算概念相似度,在一定程度上和人的直观更加符合的结果.

A Concept Similarity Computing Method Based on Hownet

Concept similarity computing,is the basic of sentence computing,this paper,based on the relationship of sememe in Hownet,consider that two concepts’ independent sememes plays a decisive role in similarity computing;the closer the independent sememes are,the smaller gap between two sememes is,the greater their similarity.This paper proposes a new method on similarity computation of concept which is based on Hownet,The method defines a similarity computation of concept among Hownet’s sememes.According to experiments,demonstrate this method have improved performance in measuring concept similarity.