A Study On The Automatic Classification Of Cultural Resources Of Minority Nationalities

21stcentury is an era of information explosion. With the growing of our national awareness of protecting minority culture resource and mass data getting into internet,even minority cultural knowledge which has lower updating rate has large quantities of accumulation. And batch after batch of national cultural websites have been set up,even lots of national cultural data bases have been built. Then how do people obtain information they needed quickly for large numbers of minority culture resource. The best way to do this is classifying the information. And this is also the premise of quick and accurate searching. But classifying information by human wastes time and energy and has low accuracy. So the automatic classification technology appeared for this. The research about automatic classification technology has a history of over fifty years. Some mature theories and methods also appeared. Whereas the application that uses this technology on the classification of minority culture resource is very few. So this paper combined the features of minority culture resource based on study for automatic classification technology and constructed a minority culture automatic classification system. Details are as follows:(1)This paper studied the present situation of automatic text categorization,summarized the process of it and made thorough research to every step according to the process. These work laid a solid theory foundation for the construction of classification system.(2)This paper analyzed and summarized characteristics of minority culture resource and constructed a minority culture resource knowledge base based on domain ontology by using the advantage of domain ontology and uniting the theory of constructing knowledge base. It provided supports and data source for construction of minority culture automatic classification system.(3)This paper also designed an automatic classification system to classify the gathered minority culture resource automatically based on theoretical research and construction of knowledge base. Then combining with the system of hierarchical classification, it improved the planar classification model and put forward the hierarchical classification model. It concluded that hierarchical classification modelhave advantages in precision ratio, recall ratio and value of F1 by experiments.