Research On Tagging Of Part-of-speech Subcategory In Modern Chinese

Research on tagging of part-of-speech subcategory in modem Chinese is the foundation for die research on NLP(Natural Language. Processing) based on corpus, also a new project for further research.It references the international methods about the auto-classifying and tagging verb subcategories,and analyses the internal research situation about some related fields, and investigates some resources,such as the subcategory system, the part-of-speech tagging method and corpus etc. It proposes a statistics integrated rules tagging model for part-of-speech subcategory and introduces vocabulary VSM and fuzzy set theory into this field.Experiments respectively adopt the tagging model based on part-of-speech information and vocabulary VSM methods through comparing the traditional tagging methods. Then combines the two techniques to build the tagging model of part-of-speech subcategory.And it improves the tagging model by two ways. It adopts the hierachical clustering in vocabulary VSM model because of its special function, on the other hand enriches the subcategory tagging information by rules, it can decrease me data sparse problem, and introduces the confidence intervals into the model for the selection of priority between statistics and rules.