用于小样本模式识别的最小平方支持向量机  

Application of the least square support vector machine in recognition of small samples

作　　者：马海兴[1] 郑璐石[1] 

机构地区：[1]宁夏大学物理与电气信息工程学院,宁夏银川750021

出　　处：《宁夏工程技术》2006年第4期386-388,共3页Ningxia Engineering Technology

摘　　要：为了进一步提高支持向量机分类器的推广性能,采用最小二乘原理,形成最小平方支持向量机,编制了相应的MATLAB程序,并将其应用于小样本模式识别中.仿真实验结果表明,最小平方支持向量机分类器在小样本模式识别中,有着优良的推广性能.In order to raise generalization performance of support vector machine classifier, this paper introduced the least square principle to obtain the least square support vector machine, programing in MATLAB, and then applying it to recognition of small samples. Simulation experiment result indicated that the least square support vector machine classifier has excellent generalization performance in recognition of small samples.

关 键 词：统计学习理论 最小平方 支持向量机 分类器 核函数 

分 类 号：O236[理学—运筹学与控制论]