Text Classification Based on Central Vector Method and Universal Gravitation

This paper presents a text categorization method (CGMV) that combines the central vector model with the gravitation mode. The intra-class information entropy and inter-class boundary confidence degree are introduced to solve the problem that the classification accuracy is degraded due to the heterogeneity of inter-class distribution and the sparse distribution of individual samples in the traditional model. Experiments show that CGMV algorithm improves the efficiency of text categorization.