一种基于《知网》的词语相似度计算方法

Computing Method of Word Similarity Based on the HowNet

随着信息检索技术的发展,自然语言处理成为研究热点。词语的相似度技术在信息检索、机器翻译等方面都有着重要的作用。文章以《知网》作为知识库,通过考虑义原地位的不同,采用不同的计算方法,综合得出词语相似度,经过实验验证,此算法与传统算法相比,合理性更胜一筹。

With the development of information retrieval technology, Natural Language Processing has become a research hotspot. Word similarity technology plays an important role in information retrieval, machine translation and so on. Taking ＂HowNet＂ as a knowledge base, by considering the different position of original meaning and different computing methods, the paper concluded the similarity of words. Compared with the traditional computation, this method is verified more rational through the experiment.