期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

双正则化参数SVM的不同实验结果

Different Simulation Results by SVM with Double Regularization Parameters
下载PDF
导出
摘要 Vapnik等人在统计学习理论和结构风险最小化的基础上提出了支持向量机算法(SVM)。该算法在高维模式识别、非线性、小样本等方面有较好的性能,已在许多领域取得了成功的应用。双正则化参数支持向量机的出现,更进一步提高了对于两类问题的研究成果。而利用样本数量来调整两个正则化参数C+和C-的比值,这一方法也得到了较为广泛的应用和认可,特别是在两类数据集中,即使这一理论没有经过严格的数学证明。文章以USPS数据库为基础,对之进行了大量的仿真实验,结果表明这个被广为接受的观点是不成立的。 Support Vector Machines （SVM） was proposed by Vapnik ct. al based on Statistical Learning Theory and Structural Risk Minimization. SVM has outstanding performance in high dimensional pattern recognition, nonlinearity, small sample data, etc. SVM has been successfully put into many areas. The research achievements have been improved by SVM with double regularization parameters. The technique of adjusting the ratio of two regularization parameters, using sample quantities has been accepted wildly, hut the theory has not been verii^cd in a serious mathematical proof. The ex- porlmental results based on USPS show that the existed theory is not correct in any conditions.
作者 姚程宽
机构地区 安庆医药高等专科学校公共基础部
出处 《成都师范学院学报》 2014年第3期118-121,共4页 Journal of Chengdu Normal University
关键词 统计学习 VC维 支持向量机 statistical learning VC dimension support vector machines
分类号 TP311 [自动化与计算机技术—计算机软件与理论]
  • 相关文献

参考文献12

  • 1Vapnik VN, Statistical Learning Theory [ M ]. NewYork: Wiley, 1998.
  • 2Boser BE, Guyon IM , Vapnik VN, A training algorithm for optimal margin classifiers[ C]//In 5th Conference on Computational Learning Theory , Pittsburgh, USA , 1992 : 144 - 152.
  • 3Cortes C, Vapnik VN, Support vector networks [ J]. Machine Learning, 1995, 20 (3) : 273 -297.
  • 4Osuna E, Freund R and Girosi F, Support Vector Machines : Training and Application [ R ]. AIM1602, MIT AI Lab 1997.
  • 5Chew HG, Crisp DJ, Bogner RE and Lira CC, Target Detection using Support Vector Machines with Training Size Biasing[ C]. 6th International Conference on Control, Automation, Robotics and Vision, 2000.
  • 6Seholkopf B, Smola A, Williamson R and Bartlett PL, New Support Vector Algorithms [ C ]. Neural Compution , 2000 (12) :1207 - 1245.
  • 7Chew HG , Bogner BE and Lim CC, Dual v - Support Vector Machine With Error Rate And Training Size Bia- sing, Digital Object Identifier [ C ]// International Conference on Acoustics, Speech, and Signal Processing, 2001 : 1269 - 1272.
  • 8Xu P and Chan A, Support Vector Machine for Multi - class Signal Classification with Unbalanced Samples E C ]// International Joint Conference on Neural Net- works, 2003:1116 - 1119.
  • 9Kim H and Park H, Protein secondary structure predic- tion based on an improved support vector machines ap- proach[ J ]. Protein Engineering,2003, 16 (8) : 553 - 560.
  • 10Morik K, Brochhausen P and Joaehims T, Combining statistical learning with a knowledge - based approach-A case study in intensive care monitofing[C]. 16th International Conference on Machine learning , 1999.

共引文献4

成都师范学院学报
2014年 第3期

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部