框架网络与汉语信息处理  

FrameNet and Information Processing in Chinese

作　　者：陶明忠[1] 马玉蕾[2] 

机构地区：[1]上海师范大学对外汉语学院,上海200234 [2]上海交通大学外语学院,上海200240

出　　处：《语言文字应用》2007年第4期120-128,共9页Applied Linguistics

摘　　要：框架网络的特点是它系统地描写了我们理解词汇和进行推理时所依赖的情景,描写了情景中语义角色的句法实现方式,对语义角色的描写比格语法更清晰、细致、系统。本文用汉语实例说明,构建汉语框架网络对解决汉语信息处理中的语义问题具有重要意义,并提出在语义角色自动识别方面,除了运用机器学习方法之外,基于规则的语义特征核查方法很值得尝试。One of the most important characteristics of FrameNet is that it provides a systematic description of the situations which we draw on when we try to understand the words or try to make inferences in sentences. It also demonstrates the ways in which the semantic roles are realized in syntactic structures. It provides much finer and more systematic description of semantic roles than Case Grammar, hence better meets the needs of information processing. The paper proposes that Chinese FrameNet be established in line with the underlying structure of English FrameNet. The paper also discusses the prospects and methods in which Chinese FrameNet is applied to information processing, and suggests that, apart from machine learning technique, the semantic features of frame elements be utilized as a plausible way to automatically identify and assign semantic role tags to the arguments of predicators in sentences.

关 键 词：框架网络 语义角色 信息处理 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]