The Study And Application Of Support Vector Machines

Support Vector Machine or SVM is new machine learning technique developed from the middle of 1990s. Being different from traditional Neural Network or NN, NN is based on traditional statistics, which provides conclusion only for the situation where sample size is tending to infinity, while SVM is based on Statistical Learning Theory or SLT, which is a small-sample statistics and concerns mainly the statistic principles when sample are limited, especially the properties of learning procedure. SLT provides us a new framework for the general learning problem. A large number of experiments have shown that SVM has not only simple structure, but also better performances, especially better generalization ability. SVM can solve small-sample learning problem better, and it is a completely new Neural Network technique. Currently, SVM is becoming a new hot area in the field of machine learning in the world. This paper has the first times systematically studied SVM deep.The main contribution of the dissertation can be summarized as follows:Firstly, chapter 1 has introduced the fact that naissance of SVM because of the development NN and some shortages of the current SVM. And we have brought forward some shortages of the current SVM. Chapter 2 has systematically discussed machine learning problem, which is the basic of SVM, with Statistical Learning Theory or SLT. Secondly, chapter 3 has educed the optimal hyperplane from pattern recognition. According to the different sample set, we have been on discussion, using Lagrangian Multiplier Technique or LMT in the optimal theory, SLT and function analysis,then we get the decision function and SVM with the corresponding different sample set.Thirdly, for improving generalization ability, application ability and recognition speed of SVM, we have used Fuzzy Set Theory (FST) and Rough Set Theory to study SVM deep, and integrated them into SVM, constructed FSSVM (Fuzzy Set SVM) and SVM based on Rough Set Theory, and extended performances of SVM in the chapter 4,5.Finally, chapter 6 is two applications on idea and approach of SVM. One is regression analysis, the other is PCA (Principal Components Analysis). We have improved their ability of anglicizing and dealing with data, and simplified procedure of processing them.