A Novel Knowledge Discovery Method for Chinese Architectural Document

Aiming at the problem of the traditional feature selection that threshold filtering loses a lot of effective architectural information and to improve the precise of Chinese architectural document classification, a new algorithm based on rough set and C4.5Bagging is proposed for Chinese architectural document categorization. Firstly the cores of attribute are found by discernibility matrix and one of the cores is regarded as the start point. Then attributes´ significance and dependency are used as the heuristic information to do feature selection. Finally the c4.5bagging is designed to architectural document classifier. The experimental results show that the novel method is not only easy to implement but can effectively reduce the dimensional space, and improve the accuracy of classification.