Analysis of Sentence Pattern of Simple Declarative Sentence in Modern Mongolian Language Based on Corpus

In this paper, on the basis of the traditional Mongolian grammar research using the theories and methods of computational linguistics, from modern Mongolian Corpus 1,000,000 words and sentences, select the primary language part of the simple declarative sentence (990), focuses on the observation sentence structure, types of sentences, the use of a variety of sentence patterns and frequency, a simple declarative sentence analysis of the modern Mongolian corpus. The library contains the boundaries of the sentence elements, the order of sentence elements in which the level of sentence elements, the elements of the sentence elements, sentence, sentence cases and other aspects of information, on this basis, build a simple statement of the modern Mongolian sentence type library. Analysis based on a simple declarative sentence of the modern Mongolian Corpus summarized five sentence elements of the subject, predicate, attribute, object, adverbial decision rule, the establishment of the modern Mongolian simple declarative sentence elements identification rule set. Which, subject identification rules 72, 94 predicate identification rules, object identification rules 73, 50 the attributive identification rules, the adverbial identification rules 69, 26 non-sentence elements identification rules. Based on the rule set, developed a simple declarative sentence elements of the modern Mongolian automatic annotation process, attempt to do the automatic annotation of sentence elements in modern Mongolian Corpus 1,000,000 words and simple declarative sentences (6300), the accuracy rate reached a preliminary by 85.6%. The full text of Chapter 5. Chapter 1, Introduction, brief introduction to the modern Mongolian research significance of the simple declarative sentence, and studies based on research methods, and are expected to reach goals. Chapter 2, a brief description of the theoretical background of this paper involved. Chapter 3 of the sentence analysis methods and modern Mongolian analysis of simple declarative sentences constitute a form of Corpus. Chapter 4 describes the the modern Mongolian simple declarative sentence library construction process and the actual content. Chapter 5 describes the procedure for the establishment of a modern Mongolian simple declarative sentence the component identification rules set method, as well as the development of modern Mongolian simple declarative sentence elements automatically marked. Conclusion concludes the paper and the outlook for the future work of the modern Mongolian sentence. In summary, this study belong to the Mongolian information processing areas of basic research, the modern Mongolian simple declarative sentence libraries and sentence elements automatically marked Mongolian information processing infrastructure programs are indispensable knowledge resources and tools. From a macro point of view, the analysis of a sentence, marked sentence elements, to confirm the sentence is machine translation, the only way of important application areas such as natural language understanding; micro ingredient, they provide knowledge resources to check the sentence with the relationship, the integrity of the judge sentence structure will play a positive role.