The Annotation Study In Modern Uyghur Language Corpus Based On Dependency Grammar

Computational linguistics needs Linguistic theories, a set of feasible linguistic theory is necessary for computer to handle human language. Dependency grammar with a lexical nature, integrity, relational, single-level and other characteristics of modern linguistic, has been widely used in computer linguistics, and contributed to the development of the theory of dependency grammar. The key point of the dependency grammar is the relationship between the words in the sentence, the verb as core words directly reflect the core elements of the sentence, and it is easier to transform from syntactic level to semantic level, more appropriate to deal with free word`s order languages, and Uyghur language as a adhesive language with free word`s order, can get a better explanation in dependency grammar. In recent years, dependency Treebank had been established in English, Czech, Russian, Japanese, Chinese, Turkish, Arabic and other languages. As a language of 740 million people use, establish the Uyghur dependency Treebank is extremely urgent. The first step in building dependency Treebank is to lay dawn a complete dependency annotation system as standard. In this paper, discuss how to annotate a Uyghur corpus through dependency relations, the dependency annotation standard. And describes establish of Uyghur dependency Treebank(UDT). In this paper, the following aspects were studied:1) Introduce the Uyghur dependency grammar annotation system which develop by the relationship between the words in Uyghur;2) A detailed describe of the dependency system through the phrases;3) Using statistical method,the discuss annotation distance, occurrence quantity of the relations, and analysis the characteristic of words in each relations;4) Existing problems and summary.