基于句法与语义特征分析的朋友关系挖掘

Mining friend relationship based on syntactic and semantic feature analysis

本文基于句法与语义特征分析提出一种朋友关系挖掘方法。首先,提出综合知网和同义词词林的词语语义相似度计算方法,从两者库中提取朋友关系描述词,得到朋友关系描述词库;其次,在处理微博文本数据时运用核心谓词、依存句法以及语义分析等方法,从微博文本中挖掘朋友关系;其中,核心谓词能确定朋友实体对,依存句法能描述前后词语间存在的关系,语义分析能进一步挖掘文本信息。实验结果证明朋友关系挖掘准确率显著提高。

This paper proposes a friend relationship mining method based on syntactic and semantic feature analysis. First, the method of calculating the semantic similarity of words in the CNKI and Synonymy Thesaurus is proposed. The friend relationship description words are extracted from two dictionaries, and the friend relationship description dictionary is obtained. Then, add core predicates, dependency syntax, and semantic analysis when working with text data, mining friend relationship from microblog texts;the core predicate can determine the entity pair;dependency syntax can describe the relationship between words before and after words;semantic analysis can further mine text information. Experimental results show that the friend relationship mining performance is significantly improved.