摘要
本文针对两种不同用途的支撑矢量机 ,分类支撑矢量机和回归支撑矢量机 ,分别证明了它们的一些几何性质 ,从这些性质出发讨论了这两种支撑矢量机对新增样本的推广能力 ,新增样本对支撑矢量 ,非支撑矢量的影响以及新增样本本身的一些特点 ,得到了一些非常有价值的结论 .从这些结论可以看出支撑矢量机对新增样本具有良好的推广能力 ,即对新增样本的良好的包容性和适应性 ,并且支撑矢量机是一种可积累的学习模型 .
Some geometry of Support Vector Machines for classification and regression is described and proven.And then the generalization performance of SVMs on newly added samples is discussed.Through the analysis of the property of newly added samples and the effect of them on support vectors and non support vectors,some valuable results are presented.These enable us to conclude that SVM has a good compatibility,adaptability and generalization performance for newly added samples and is a hereditable learning model.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2001年第5期590-594,共5页
Acta Electronica Sinica
基金
国家"8 63"项目! (No.863 30 6 0 6 0 6 1 )
教育部博士点基金
关键词
分类支撑矢量机
回归支撑矢量机
学习机
support vector classification
support vector regression
learning machines
KKT conditions
hereditability