The Research Of Semantic Role Statistical Analysis Of Mongolian Noun Phrase

This thesis studies the Mongolian noun phrase acted as semantic role by using statistical analysis. The methodology includes the phrase structure grammar, case grammar and valence grammar of modern linguistic theories. The research conducted semantic role labeling for5107simple sentences of "one million words level of modern Mongolian corpus"," GEERLECHAOKETU complete works"(volume1-4)," Inner Mongolia daily"(corpus of10days in2006), statistical analyzed7646noun phrases as semantic role and summed up the semantic role recognition rules of noun phrase. Semantic role recognition of Mongolian noun phrase is one of the main content of Mongolian semantic role labeling.This dissertation consists of five chapters:The introduction is mainly about the meaning, research status, methodology, data resources and tagging-set.Chapter one introduces the selected sentences, proofreading results of words and phrases, part of speech, case-tagging related terms and phrases semantic role labeling.Chapter two discuss the construction of analysis base of Mongolian noun phrase semantic role and mainly studies the statistical related items, such as morphology of their noun phrase acting as semantic role, classification of center noun and dominating verb acting as semantic role and so on.Chapter three summarizes813noun phrase semantic role recognition rules based on the analysis of the complex features.The last chapter gives a brief summary and present research of further study.