A complementary hybrid classification algorithm based on Webtext

Lili Xing, Bing Zhang, Yuhong Lu, Zhong Li

COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12B) 258-263

Institute of Disaster Prevention, Information Department, Yanjiao 065201, China

In view of the insufficiency of existing weight computation methods and SVM algorithm, a weight computation method of variable precision rough set based on Web text and a complementary hybrid classification algorithm are proposed; In the hybrid classification algorithm, the rough set is used as a front-end processor of SVM, the traditional SVM is optimized from classification efficiency and precision through the reduction theory and weight computation method proposed in this paper. The experimental results show that the reduced and weighted data are classified using SVM, and then the performance of classification is further guaranteed.