The Construction Of The Knowledge-Base For Mongolian Conjunctive Form

In order to meet the requirements of Mongolian information processing, we focus on the research of Mongolian conjunctive form, including the structure, function, meaning as well as its automatic recognition from the perspective of computational linguistic. In this study, we develop the Knowledge-base of Mongolian Conjunctive Form which contains the information of morphology, syntax, semantics and usage. Based on the above knowledge-base, we also develop the Rule Database of Mongolian Conjunctive Form. This thesis consists of five chapters.Chapter 1 introduces the research object, significance, background, methods and materials. Based on the previous work on the Mongolian conjunctive form, we explain the naming, definition, classification, range of Mongolian conjunctive form as well as other related issues.Chapter 2 explains the attributes in the database such as the structure, function and position which comes from the analysis of the corpus. The Mongolian conjunctive form is divided into two classes:composite structure and anaphoric structure. We focus on two aspects:the position of the conjunctive form in a sentence and the relationship between the conjunction itself and its connected components. While analyzing the function of the conjunctive form, we record the number of occurrences in the corpus, the former connected components and the conjunction functions and so on. The conjunctive forms in Mongolian language have nine types of relations:coordination progression selection,adversative,supposition,opposition,going-ahead,cause-effect and hooking.Chapter 3 analyzes the structure and the function of the conjunction form based on its usage in the corpus and the traditional Mongolian linguistics. We set the attribute fields for Knowledge-base of Mongolian Conjunctive Form and fill in the fields on the basis of its structure and function. The attributes are not based on syntactic information but on semantic information. That is the difference with the Mongolian grammar information dictionary.Chapter 4 introduces the construction of the Rule Database of Mongolian Conjunctive Form which is based on the phrase structure grammar and presented in the text form. We divide the Mongolian conjunction rules into two classes:the rule for composite structure and the rule for anaphoric structure.Chapter 5 elaborates the development of the Automatic Recognition program for Mongolian conjunctive form, including the system architecture and data flow chart and so on. The performance of the program is also presented at the end of this chapter.