一种基于义素的网页信息项语义匹配方法研究

A Sememe Based Semantic Matching of Web Information Item

本文提出了一种改进的基于语义的义素相似度,并从理论上分析参数β值的影响效果。在这个基础上,提出一种基于义素的词相似度,从语义上去匹配新名词和旧名词。在基于义素的词相似度基础上,提出一种网页信息项的语义匹配方法,来识别网页信息项的类别。实验结果表明,基于义素相似度的网页信息项语义匹配方法具有较好的匹配效果。

Because some new words will be created in the development of human knowledge and the expression forms of words may be various, the word matching method is needed to research in order to adapt the changes from the point of the semantic features of words. This paper proposes an improved sememe similarity, which is based on se- mantic features of words. To analyze the effect of the coefficient β, this paper deducts three theorems in theory and then educes that the value of the coefficient β must be set in a range, which is to say that the coefficient β cannot be set much little value or much large value. Based on the sememe similarity, a word similarity is put forward to match new words with old words from the view of the semantic features. To adapt the Web page information, which in- cludes the human knowledge and has various expression forms, a novel semantic matching method is proposed to iden- tify the class of the Web information items. The semantic matching method is based on both the word similarity and the sememe similarity. As the experiment results show, the sememe based semantic matching method gains higher ac- curacy to identify the class of the Web information items.