用球结构的支持向量机解决多分类问题

支持向量机是从统计学习理论中导出的，从研究二分类开始，发展至今，虽然提出了很多多类别分类的相关算法，但都各有不足之处。提出基于球结构的支持向量算法，适用于规模比较庞大的多类别问题，并对其性质进行了讨论。

Solving the Problem of Multi-class Pattern Recognition with Sphere-structured Support Vector Machines

Support vector machines (SVMs) are learning algorithms derived from statistical learning theory. The SVMs approach was originally developed to solve binary classification problems. There are some methods to solve multi_class classification problems, such as one_against_rest, one_against_one, all_together and so on. But the computing time of all these methods are too long to solve large_scale problems. In this paper SVMs architectures for multi_class problems are discussed. Furthermore we provide a new algorithm called sphere_structured SVMs to solve the multi_class problem. We show the algorithm in detail and analyze its characteristics. Not only is the number of convex quadratic programming problems in sphere_structured SVMs small, but also the number of variables in each programming is the least. The computing time of classification is reduced. Otherwise, the characteristics of sphere_structured SVMs make it easy for data to expand.