【作者】 白宇； 于水； 叶娜； 蔡东风； 任晓娜；

【Author】 BAI Yu,YU Shui,YE Na,CAI Dongfeng,REN Xiaona (Knowledge Engineering Center,Shenyang Institute of Aeronautical Engineering,Shenyang,Liaoning 110034)

【机构】 沈阳航空工业学院知识工程中心；

【摘要】 词语相似度计算是机器翻译、信息检索等自然语言处理领域的关键问题之一。传统的词语相似度计算方法,未能很好地考虑上下文信息对词语语义的约束,从而不能对语境变换带来的词语间相似度的差异进行有效的区分。本文引入模糊数学中隶属函数的概念计算词语上下文信息的模糊重要度,并结合基于知网的语义相似度计算方法,提出一种基于语境的词语相似度计算方法。实验表明,该算法可以根据语境有效的区分语义相近的词语。更多还原

【Abstract】 Word similarity computation is one of the key issues in natural language processing fields,such as machine translation,information retrieval etc.As traditional methods ignored the context information of the word,which can not distinguished in effect the difference of the words similarity when the context information changed.This paper presented an approach for word similarity computation based on the context information,which employed the fuzzy membership functions to compute the fuzzy significance of the words,and combined a method of word similarity calculation using HowNet.The experimental results indicate that our approach distinguish the words with similar semantic by the context information effectively.更多还原