计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (17): 74-76.

• 学术探讨 • 上一篇    下一篇

回归支持向量机SMO算法的改进

许建潮,张玉石   

  1. 长春工业大学 计算机科学与工程学院,长春 130012
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-06-11 发布日期:2007-06-11
  • 通讯作者: 许建潮

Improvement of SMO algorithm for SVM regression

XU Jian-chao,ZHANG Yu-shi   

  1. School of Computer Science & Engineering,Changchun University of Technology,Changchun 130012,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-06-11 Published:2007-06-11
  • Contact: XU Jian-chao

摘要: 在Smola 和Sch?觟lkopf的SMO算法中,由于使用了单一的极限值而使得算法的效果没有完全表现出来。使用KKT条件来检验二次规划问题,使用两个极限参量来对回归SMO算法进行改进。通过对比实验,这一改进算法在执行速度上表现出了非常好的性能。

关键词: 支持向量机, 回归, 序列最小优化

Abstract: Smola and Sch?觟lkopf’s SMO algorithm is inefficiency sometimes,because the algorithm uses a single threshold value.In this paper the KKT conditions is used to check up the dual problem and two threshold parameters are employed to derive modifications of SMO for regression.Through the contrasted experiment,the modified algorithm performs very well about capability on executing speed.

Key words: Support Vector Machine(SVM), regression, Sequential Minimal Optimization(SMO)