名词隐喻的计算研究及识别实验  

The Computational Research of Noun Metaphor and Recognition Experiment

作　　者：王治敏[1] 

机构地区：[1]北京语言大学汉语学院,100083

出　　处：《语言教学与研究》2008年第2期68-74,共7页Language Teaching and Linguistic Studies

基　　金：国家973项目(2004CB318102);北京语言大学一般项目(06YB06)

摘　　要：本文通过短语结构限定的方式,全面地考察汉语名词性隐喻的分布和构成特点,总结和发现名词性隐喻的表达规律,同时在大规模语料考察的基础上,尝试规则约束和机器学习方法,探讨了汉语名词性隐喻的理解和识别。实验证明,机器学习方法在汉语名词隐喻识别中取得了较为满意的效果。This paper aims to fully examine the distribution and structure of n ＋ n metaphors in Chinese, which contain noun phrase structure. We also explore the understandings and recognitions of noun metaphors based on the survey of the language corpus and with the help of rule restriction and the machine learning techniques. Our experiments show that the recognition results of the ma- chine learning method is satisfied. The plan for further research includes establishing the similarity inference mechanism for noun metaphor to improve recognition performance.

关 键 词：源域 目标域 隐喻识别 最大熵 

分 类 号：H15[语言文字—汉语]