Research On Semantic Role Of Agent And Experiencer Of Mongolian Simple Sentence Based On Corpus

Mongolian semantic role labeling is a realization way of shallow semantic parsing and entry from shallow to in-depth study.In this Papers, on the basis of5087simple verb predicate sentence we extracted the features of agent and experiencer which collected from the Mongolian dependency tree-bank, and describe the automatic recognition rules and develop the program.Therefore, the performance of the system is improved effectively, for the system gets its result with87.62%recall and93.0%precision.This paper consists of an introduction, the main parts and conclusion.The introduction part, main aspect, including the meaning, preview, objectives, goal, methodology and data sources is introduced.The first chapter, classification of semantic roles, tagging-set and process using dependency grammar is discussed.The second chapter, features of Mongolian agent and experiencer according to the their dependent relation, part of speech, semantic classification, and collocation based on labeled dependency data is summarized. The third chapter, annotation of Mongolian agent and experiencer based on the dependency grammar is established.The fourth chapter, we annotated the Mongolian agent and experiencer, and also discussed the annotation results and puts forward the existing problems and solutions.At the last section, brief summary together with the future directions are indicated.