Shallow Parsing On Tibetan Textbook Of Junior Middle School

In the area of natural language processing, full parsing is a key point as well as a difficult point. Full parsing is not believed to be solved thoroughly based on the present study. But partial parsing, i.e. chunk parsing, can be used well to not only process natural language sentences but also lower the difficulty of analysis.Then on the basis of the previous definition of chunk, there are six types of chunk defined, which depends on syntactic marks in Tibetan textbook of junior middle school and grammar rules of modern Tibetan. Carried out the research of the following aspects:blocks of type, statistics of frequency and distribution and blocks the statistical co-occurrence relations, Intuitive evaluation of the structural composition and content of the junior high school of Tibetan language materials constitute.. And with the group block distribution reflects the value of the blocks.One part of the shallow parsing system is studied in this thesis, i.e. the chunk statistics and analysis of Tibetan textbook of junior middle school besides the difficulty. Two parts of chunk statistics are not only the general chunk type statistics but also the chunk frequency distribution and co-occurrence relations among chunk. As a result, the scientific difficulty of textbook is analyzed according to the statistics and some problems and shortage of textbook are found out. Furthermore, it can give some reasonable suggestions on designing textbook through data analysis.In this paper, the chunk simplifies sentence structure and improves the entire function of machine translation. It offers an idea for evaluating Tibetan textbook of junior middle school objectively, the foundation for improving the compiling quality of Tibetan textbook. Meanwhile, the result can be used in other areas of natural language processing, like information retrieval and text categorization.