计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (26): 32-34.

• 研究、探讨 • 上一篇    下一篇

线性规划ν-支持向量机的牛顿法

宋 杰   

  1. 韶关学院 数学与信息科学学院,广东 韶关 512005
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-09-11 发布日期:2011-09-11

Newton algrithm for ν-support vector machine based on linear programming

SONG Jie   

  1. College of Mathematics and Information Sciences,Shaoguan University,Shaoguan,Guangdong 512005,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-09-11 Published:2011-09-11

摘要: 基于线性规划的ν-支持向量机(ν-LPSVM)是在基于二次规划的ν-支持向量机(ν-QPSVM)的基础上提出的。ν-LPSVM和ν-QPSVM模型中的参数ν都可以控制支持向量的数目和误差,但ν-LPSVM的模型更为简单,应用前景更广。讨论了这种新型支持向量机的线性规划问题的最小2-范数解,在此基础上给出了一个快速、有限步终止的牛顿算法。数值实验表明,ν-LPSVM的牛顿算法快速而且有效。

关键词: 支持向量机, ν-支持向量机, 线性规划, 牛顿算法

Abstract: The ν-Support Vector Machine based on Linear Programming (ν-LPSVM) has been proposed following ν-support Vector Machine based on Quadratic Programming (ν-QPSVM).The parameter ν can also control the numbers of support vectors and errors.How ever,the formulation of ν-LPSVM is much simpler,and its application is more prospective.The least 2-norm solution of the linear programming formulation of the new SVM is discussed.Based on it,a fast Newton algorithm is proposed,which is terminated finitely.The numerical tests show that the Newton algorithm of ν-LPSVM is fast and effective.

Key words: Support Vector Machine(SVM), ν-Support Vector Machine(ν-SVM), Linear Programming(LP), Newton algorithm