| With the development of the technology in network and communication, mass information crowd in on. How to effectively select what we need becomes more outstanding than ever. Data mining as a processing technique is popular to be used to meet the need.The classification technology, as an important aspect of data mining, has always been the concern of the researchers. Many good solutions have been created during the classification algorithm research process. The support vector machine, which is the focus of this paper, is a very effective classification method. It is thought of a new generation of learning machine based on the statistical learning theory. The main advantage of SVM is that it can serve better in the processing of small-sample learning problems by the replacement of experiential risk minimization by structural risk minimization. Moreover, SVM can treat a nonlinear learning problem as a linear learning problem since it maps the original data into the kernel space in which we only solve the linear learning problem.At the beginning the concepts, the research state in the world, the application, the process and the basic method of data mining are addressed. And next this paper studies the basic knowledge of statistical learning theory. Then the paper stresses to introduce SVM, involved of the development history, the state so far, main concepts and the content of research. Na?ve SVM is only able to deal with binary classification. In this thesis, after discussed the current multi-class SVMs, a novel multi-class SVM classifier based on the relativity separability measure is proposed. And the numerical and applicable results illustrate that the technique is feasible and effective in the end. Otherwise by analyzing the characteristics of support vectors and the processing of incremental learning, a new algorithm of incremental learning method is presented in this paper. Experimental results have shown that, with keeping the testing accuracy, the new method discards useless samples and reduces the training time.The research of SVM in domestic just limits in theory so far and the application do start just now. SVM is a technology for application, so there are still much knowledge and experience that need to be found in practice. |
