基于知网和知识图的汉语词语语义相似度算法

A Chinese Word Semantic Similarity Algorithm Based on HowNet and Knowledge Graph

汉语词语间语义相似是词语间的基本关系之一,文章提出了一种基于知网和知识图的词语语义相似度计算的方法,通过改进传统的知识图表示方式,根据知网中概念项的抽取结果对词语的义项进行表示,用词图的相似度来表示相应词语的语义相似度。实验结果表明该算法对词语间语义相似度计算是有效的。

Semantic similarity is one of the basic relations between Chinese words.In this paper,the words semantic similarity based on HowNet and knowledge graph is presented,the traditional knowledge graph have been improved,expressed the concept of word according to the extracted results of DEF in HowNet,the similarity of word graph was respond to the similarity of words.The experiment result proves this algorithm is effective for word semantic similarity computation.