词语语义相关度的计算,一种比较常用的方法是使用分类体系的语义词典,而国内外学者已经提出了多种基于语义相关的度量方法。这些方法对于词典和语言环境的依赖性是一个值得研究的问题。本文汇总了多种基于语义词典的方法,全面地概括分析了这类方法的特点。基于哈尔滨工业大学信息检索实验室提供的《同义词词林》扩展版,本文在真人单词对相关度判断实验中比较了多种方法的效果,从而找出了《同义词词林》扩展版中的较好方法。
To compute the semantic relatedness of words, a frequently - used method is to use the classified semantic dictionary. Scholars both at home and abroad have already proposed multiple measuring methods based on semantic relatedness. The dependence of these methods on dictionaries and context is a problem worthy to be stud- ied, This paper collects multiple semantic dictionary - based methods, and sums up the characters of these methods in an all - round way. Based on the Tongyici Cilin (Extended) of the IR Lab of HIT, this paper compares the ef- fectiveness of multiple methods in the application of the human couple words in relatedness judgment experiment, thereby finding the better method in the Tongyici Cilin (Extended).