Automatic Document Classification of Digital Library via Kernel Method

The issue of the automatic classification is primary importance for scientific databases and digital libraries. A sophisticated classification strategy renders searching more effective and assists the users in locating similar relevant items. We propose a new robust document classification algorithm using a multiscale kernel, according to the wavelet frame theory and kernel function nature in this paper. The hierarchical tree classifiers have been proposed for multi-classification. The choice of kernel parameters is simplified and fast in the proposed model. Compared with the other kernel function classifier for document classification methods in literature, proposed method is more efficient in both training and classification. Experimental results on digital library document classification demonstrate that our method is not only significantly more efficient but also achieves high classification rate. The experimental result also shows that multiscale classifier is the validity and the usability.