Performance Analysis of Chinese Webpage Categorizing Algorithm Based on Support Vector Machines (SVM)

Abstract

Categorizing web automatically for users is a key technique of information society, and the key point of this technique is web training and categorization. This paper researches one of the important algorithm in this field—support vector machines (SVM). By analyzing and simulating 4 kinds of kernel function and 3 ways of feature selection, polynomial kernel function and document frequency is chosen for the best way in SVM algorithm. Meanwhile, pre-process algorithm is given in this paper in order to improve the efficiency of categorization. By simulation, importing pre-process method to SVM enhances the capability of the web categorization both in precision and time-consumption.