现代维吾尔语依存语法语料库标注研究

发布时间：2018-04-21 10:20

本文选题：依存语法 + 标注规范　；参考：《新疆大学》2016年硕士论文

【摘要】：计算机语言学需要语言学理论,让计算机处理人类语言,需要一套切实可行的语言学理论。依存语法具备了现代语言学的词汇性、整体性、关系性、单层次性等特点,在计算机语言学中得到了广泛的运用,也助长了依存语法理论的发展。依存语法的重心在句中词语词之间的关系,以动词为核心词,直观的反映了句子的核心成分,更便于句法层面到语义层面的转换,更宜于处理自由语序的语言,而维吾尔语作为自由语序的粘着性语言,在依存语法中可以得到更好的阐释。近年来,英语、捷克语、俄语、日语、汉语、土耳其语、阿拉伯语等语言的依存语法树库相继建立,作为全球740万人通用的语言,维吾尔语依存语法树库的建立迫在眉睫。而建立依存树库的第一步是要有一套完整的依存标注体系作为标准。本文讨论了怎样用依存关系标注维吾尔语语料,即依存语法标注规范,并介绍了维吾尔语依存树库UDT1,2(Uyghur dependency Treebank)的建立。本文从以下几个方面进行了研究:(1)介绍了根据维吾尔语句子词间关系研制出的维吾尔语依存语法标注体系。(2)从句中短语出发,更详细地讨论了标注规范。(3)用统计的方法,在现已标注完的三千多个句子的基础上讨论了标注距离、标注关系出现量等问题并进行词性分析。(4)现存在的若干问题与总结。
[Abstract]:Computer linguistics needs linguistic theories, and a set of practical linguistic theories is needed for computers to deal with human languages. Dependency grammar has the characteristics of lexicality, integrity, relation and single level in modern linguistics. It has been widely used in computer linguistics, and has also contributed to the development of the theory of dependent grammar. The emphasis of dependency grammar is on the relationship between words and phrases in sentences. The verb is the core word, which directly reflects the core components of the sentence, facilitates the transformation from the syntactic level to the semantic level, and is more suitable for dealing with the language of free word order. Uyghur, as a cohesive language in free word order, can be better interpreted in dependency grammar. In recent years, the dependent grammar tree database of English, Czech, Russian, Japanese, Chinese, Turkish, Arabic and other languages has been established one after another. As a common language of 7.4 million people in the world, the establishment of Uygur dependent grammar tree is imminent. The first step to build dependency tree database is to have a complete dependency tagging system as the standard. This paper discusses how to annotate Uygur language data by dependency relation, that is, dependency syntax annotation specification, and introduces the establishment of UDT1Uyghur dependency tree library. This paper introduces the phrases of Uygur dependent grammar tagging system, which is developed according to the relationship between Uygur sentences and words, and discusses in more detail the tagging criterion. On the basis of more than 3,000 sentences that have been annotated now, this paper discusses the problems of tagging distance and the appearance of tagging relations, and analyzes the parts of speech.
【学位授予单位】：新疆大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：H215

【相似文献】