概念与属性间语义约束知识的获取方法研究  

The Constraint Knowledge Extraction from the Concept and Its Attributions

作　　者：张梅[1] 段建勇[1] 

机构地区：[1]北方工业大学,北京100144

出　　处：《语言文字应用》2012年第1期133-140,共8页Applied Linguistics

基　　金：教育部人文社科青年基金(项目编号:10YJC740132);国家语委十二五规划基金(编号:YB125-10);国家社会科学基金(项目编号:11CTQ036);国家自然科学基金(项目编号:61103112)的研究成果之一

摘　　要：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 stepping stone of semantic computing. We propose a concept representation method by its attributions and quantifying its knowledge granularity. Firstly, the seed attribution knowledge is extracted from HowNet which is a lexical semantic dictionary. Then the free attribution knowledge base is constructed and extended using the context templates. With the help of the free attribution knowledge base, we can train the constraint knowledge among the concept and its attributions and quantify their relations. Finally this method is implemented.

关 键 词：属性知识 概念语义 语义表示 

分 类 号：H030[语言文字—语言学]