Research On The Automatic Recognition Of Mongolian Political Terms

Terminology is an important carrier of knowledge,which is widely used in our life and learning,so it plays an increasingly important role.In recent years,with the rapid development of natural language information processing technology,it is possible to study terms through the method of natural language information processing,and become a noticeable research topic.The study of Mongolian term automatic recognition is of great significance to Mongolian term work,machine translation,lexicography and text classification.In this paper,according to the knowledge of politics,terminology and Mongolian related disciplines,Mongolian political terms are defined.On this basis,the paper studies the automatic recognition of Mongolian political terms based on conditional random field model.The research of term recognition based on conditional random field model mainly includes four parts: prediction preprocessing,feature selection,design feature template and result analysis.In this paper,Mongolian political terms are automatically recognized by the feature set of words,part of speech features,dictionary features and keyword features.The final recognition accuracy is 89.83%,recall rate is 69.55%,F1 value is 78.40%.The paper consists of six parts: introduction,chapter one,chapter two,chapter three,chapter four and conclusion.In the introduction part,the significance,general situation,contents,methods and steps of the research are described.The first chapter describes the basic elements of terms and political terms.In the second chapter,the research methods of automatic term recognition are introduced,and their advantages and disadvantages are discussed.In the third chapter,the specific steps of Mongolian political terms automatic recognition based on conditional random field model are described.This paper mainly introduces the corpus processing,feature selection and feature template setting.In chapter four,two groups of experiments are designed.The effects of features and training corpus on the automatic recognition of Mongolian political terms are tested and the results are analyzed.The conclusion part summarizes the content of the whole paper and briefly describes the next research content.