Support Vector Machine Algorithm Research And Application

Support Vector Machines (SVM) is a new pattern recognition technology, which is based on Statistical Learning Theory. It can solve small-sample learning problems better by using Experiential Risk Minimization in place of Structural Risk Minimization. Moreover, this theory can change the problem in non-linearity space to that in the linearity space in order to reduce the algorithm complexity by using the kernel function idea. Because it has quite perfect theoretical properties and good learning performance, SVM theory becomes the new research hotspot after the research of Artificial Nerve Net and pushes the development in machine learning theory and technology. However, SVM theory performance has been validated in many practical applications, there are still some drawbacks. For example: train speed is slow, algorithm is complex and check phase operation is large, etc. According to above problems, this dissertation mainly focuses on the following research work, especially in SVM algorithm research.At first, the basic concept of the SVM theory has been introduced and the basic theoretical properties has been gone deep into discussed. Then two classical SVM algorithms桽OR algorithm and LSSVM algorithm have been done more research and the two algorithms performance has been compared by using MNIST database.Secondly, in allusion to large-scale train set, a new simple incremental learning algorithm桽ISVM has been put forward in order to overcome the slowly train speed. This algorithm applies the small-scale matrix operation to replace the large-scale matrix operation by carefully analyzing the character of the Support Vector distributing. The experience result shows that this algorithm has effectively speeded up the train process.At last, SVM algorithm has been applied to remote sensing image classification. Compared with K Near Neighbor and Adaptive Min-distance Algorithm, the experience result presents that SVM algorithm has better classification effect. And the experience result also shows us that SVM algorithm has good application foreground in the aspect of remote sensing image classification.