Knowledge Of Ancient Architectural Heritage Protection Classification

In China's long history of 5,000 years, the first people created a brilliant architectural culture. But with the passage of time, some of the old buildings have been lost and some of them have been damaged seriously, some of them though have been repaired, but it is difficult to just showing their original appearance. In fact, much information about the old buildings has been recorded by the careful people. What we have to do is to collect the information, and classify them.Early knowledge of the ancient buildings are classified by the work of ancient architecture expert implementation manual, time-consuming effort, will be applied to text classification ancient buildings in the classification of knowledge can greatly reduce the intensity of people's work and improve utilization of information.The main performance: (1) in-depth study of the text classification involved in key technologies, such as segmentation, the vector space model, feature extraction, classification algorithms, text classification evaluation standards. (2) include achieving a Bayesian, k-Nearest Neighbor, SVM algorithm of the text classification system, feature extraction method used information gain, mutual information and Expected cross entropy,χ2 statistics, the Weight of Evidence for Text, and other characteristics of evidence to assess function. (3) compared the three classification algorithm performance of the classification, discussed the various features of the classification function to assess the pros and cons of performance on the characteristics of the various different dimension classification algorithm, as well as feature extraction method classification performance impact. (4) through analysis of the experimental results, the dissertation proposes some opinion about Text Automatic Classification and feature selection algorithm, and gives the expection for future work on them.