面向语料库标注的汉语依存体系的探讨

实现大规模真实文本的处理, 是信息化社会的迫切要求, 也是国际计算语言学界的一个战略目标目前一项迫在眉睫的任务是建立一套满足大规模真实文本处理的语言处理体系, 包括分词的标准、词的分类体系、句法体系和语义体系。其中句法体系是核心环节。本文提出并论证了依存语法是合乎大规模真实文本处理要求的句法体系, 并结合汉语的特点, 研究了汉语的依存语法, 划分了种依存关系。最后简要讨论了依存语法的一些应用

Approach to the Chinese Dependency Formalism For the Tagging of Corpus

s a strategic target of the global computational linguistic circle, the large scale authentic text processing becomes more and more significant for the modern informative society. To meet with the requirement of large scale corpus processing such as tagging, knowledge acquition and automatic analysis, an language processing formalism must be set up as soon as possible. The formalism includes the word segmentation standard, the part of speech scheme, the syntactic formalism and semantic formalism. The syntatic formalism is the key part among them. In this paper, it is presented that Dependency grammar is a suitable syntactic formalism for large scale authentic text processing, and the Dependencygrammar for Chinese is specially studied, and 44 kinds of dependency relations are defined.Finally, some application of the Dependency grammar to Chinese language processing arebriefly discussed.