Studies On The Usage Of Preposition And Conjunction In Phrase Structure Syntactic Parsing

Chinese syntactic analysis is an important topic in the field of natural language processing. Aiming at the characteristics of the Chinese, this paper adds the preposition’s usage to the output of the parser, postprocesses the output of the Stanford Parser using the usage properties of the preposition. Firstly, in order to get a higher results of preposition’usages recognition, on the basis of the existing rule-based method of preposition’usages recognition, this paper proposes an automatically recognizing method of preposition’usages based on statistics. And three statistical models, which are CRF、ME and SVM, are used to label some common preposition’usages on the tagged corpus of People’s Daily (2000.2,3,4). And the final result shows that automatic recognition of preposition’usages based on statistics is better than the result of rule-based method of preposition’usages recognition generally. And then, on the basis of the output of the Stanford parser, this paper modifies the existing results using the boundary identification received from the property characteristics of preposition’usages, and then enhances the accuracy of Chinese parser. The experiments shows that the result that has the property characteristics of usages has been improved compared to the result that has not to some extent. Finally, in order to verify the applicability of this after-processing approach, this method is used on the conjunction, and gets a good result.The main work of this paper includes:(1) According to the thoughts of building the "Trinity" knowledge base of functional words, on the basis of the correct corpus received from the proofreader for the result of preposition’usages automatic annotation based on rule, this paper achieves the automatic annotation of preposition’usage based on statistics.(2) On the basis of automatic identification of the preposition’usage, prepositional phrase boundary identification based on the usage, Stanford Parser and the Penn Chinese Treebank, this paper achieves the study on the preposition’usage in the phrase structure syntactic parsing.(3) According to the study on the preposition’usage in the phrase structure syntactic parsing, on the basis of the automatic identification of the conjunction’ usage and conjunction phrase boundary identification based on the usage, this paper achieves the study on the conjunction’usage in phrase structure syntactic parsing. Finally, this paper summarizes the content of this study, puts forward to the further work, and points out the future research directions.