Design And Implementation Of A Uyghur Chart Parser

Parsing natural language processing is a very important research. Analysis of a language, including lexical analysis, syntax analysis, semantic and pragmatic analysis of several levels, so parsing a direct impact on the process of follow-up project.This paper first established Uighur Syntax tree Library label system, the annotation system consists of two modules: Functional block tag module and component marking module. Through this tagging system, we established a rule base, and completed the ground work, we analyzed the domestic and foreign are popular methods of syntactic analysis, choose the Chart parsing, beginning with the bottom-up Chart parsing , but found that this method of low efficiency, but also prone to ambiguities, we improved the algorithm, bottom-up and top-down Chart algorithm combined to achieve a rule-based Uighur Chart parser. The test of the analyzer is in the Xinjiang Autonomous University of Multilingual Information Technology Laboratory of speech tagging corpus (XJU UPOS Corpus) on the basis of test results better.In the process of system analysis with the Chart, we find some problems, such as: the selected small-scale dictionaries and rule base, rule base often conflict between the rules and so on. Because we know that there is ambiguity of natural language, resulting in the above-mentioned problems are inevitable.Future work should continue to improve Treebank system, thus improving the accuracy of the rule base, rule base and should be more complicated grammar rules and integrity, thereby enhance the efficiency of Chart syntactic analysis of decency.