语义表示是语义计算的基石,本文提出运用属性来表征概念语义,并采用机器学习的方法来量化基于属性知识粒度的概念语义表示方法。首先从“知网”中抽取自由属性知识的种子词汇,然后采用模板的方法来扩展属性知识库;在已有自由属性知识库的基础上训练概念与属性知识的约束知识,并量化两者间的约束关系。最后对该方法进行了实现。 Semantic representation is the cornerstone of semantic computing. This paper proposes the use of attributes to represent concept semantics, and uses machine learning methods to quantify concept semantic representation based on attribute knowledge granularity. Firstly, extract the seed words of free attribute knowledge from “知知 网 ”, and then extend the attribute knowledge base by using the template method. On the basis of the existing free attribute knowledge base, train the constraint knowledge of concept and attribute knowledge, Between the constraints. Finally, this method is implemented.