面向词义消歧的词语相关度计算

Word sense disambiguation based on word semantic relevancy computation

为解决词义消歧问题,引入了语义相关度计算。研究并设计了词语相关度计算模型,即在充分考虑语义资源《知网》中概念间结构特点、概念信息量和概念释义的基础上,利用概念词与实例词间的搭配所表征的词语间强关联来进行词语相关度的计算。实验结果表明,该模型得到的语义相关度结果对于解决WSD问题提供了良好的支撑依据。

Semantic relevancy computation is used to solve the Word Sense Disambiguation （WSD） problem. A new computational model is presented and designed for word semantic relevancy between two Chinese words. With HowNet as the semantic knowledge resource, this model not only considers the structure characteristics between concepts, the information and interpretations of concepts, but also the strong relevancy of the collocation between the concept and the example. Experimental results show that the results are satisfactory and good support for solving the WSD problem.