对《知网》的补充以及类词典的建立

Complement the HowNet and Produce a Thesaurus Based on the context co-occurrence

文章基于词的上下文语义环境对词进行聚类,以对语义词典《知网》进行补充,并最终得到一个基于上下文相关词的类词典。本文提出的聚类算法是在基于共现的聚类算法的基础上。

This article makes use of a method of clustering word based on context. The result classes are used to complement the semantic thesaurus HowNet which took Mr. Dong zhengdong tens of years to compile, At last, a thesaurus based on context co-occurrence is produced. Our method improved the method based on co-occurrence data, especially improved the point that the ability of the word in co-occurrence are not equal. Based on our method, we finally get a clustering words dictionary, which represent not only the word＇s context co-occurrence but also the meaning in a sense. The dictionary is helpful to the further research of the sentence analysis and eliminating the ambiguous meaning of words.