The Study Of Web Text Mining Based On SVM

With the explosive growth of the Internet, Web has contained flood of information and rich resources. It is necessary to provide users with tools for efficient knowledge discovery on the Web, in order to improve the efficiency of information retrieval and information usage on Web. With the application of data mining to Web, Web mining has become one of the most important and flourishing fields of data mining.SVM is a new kind of promising machine learning algorithm proposed by Vapnik and his group at AT&T Bell laboratory. Because SVM has stronger theoretical foundation and better generalization performance, it becomes the new research hotspot after the research of Artificial Nerve Net and it will push the development in machine learning theory and technology.In this paper, the principle of Web mining has been introduced at first. The building of Web text mining system has been discussed in detail, including the architecture of Web text mining system and function of system module. Secondly, the basic knowledge of the statistical learning theory has been introduced and the SVM based on the theory has been gone deep into discussed. In the end SVM algorithm has been applied to Web text mining for Web text classification. An active learning method using SVM is researched. Comparing with the general SVM, it can improve the capability on the premise of keeping correctness of the classifier. It also shows us that SVM algorithm has good application foreground in the aspect of Web text mining.