期刊文献+

一种新的支持向量机多类分类方法 被引量:46

A New SVM Multiclass Classification Method
下载PDF
导出
摘要 分析了目前的支持向量机多类分类方法存在的问题以及缺点.针对以上问题及缺点,提出了基于 二叉树的支持向量机的多类分类方法,并在UCI数据库上进行了验证,取得了良好效果. The problems and defections of the existing methods of SVM multiclass classification are analyzed. To solve these problems, an SVM multiclass classification based on binary tree is put forward. In order to verify the effectiveness of this method, experiments have been made on UCI database, and the experimeatal results are satisfaoto-
出处 《信息与控制》 CSCD 北大核心 2004年第3期262-267,共6页 Information and Control
基金 国家自然科学基金资助项目(60275020)
关键词 支持向量机 分类 二叉树 迭代算法 support vector machine (SVM) classification binary tree iteration algorithm
  • 相关文献

参考文献9

  • 1Bottou L, Cortes C, Denker J. Comparison of classifier methods:a case study in handwriting digit recognition [ A]. Preceedings of the 12th IAPR International Conference on Pattern Recognition [ C ]. Jerusalem: IEEE, 1994.77 ~ 82.
  • 2Platt J C, Cristianini N, Shawe-Taylor J. Large margin DAGs for multiclass classification [ A ]. Advances in Neural Information Processing Systems [C]. 2000.547 -553.
  • 3Vapnik V. Statistical Learning Theory [ M]. New York:Wiley,1998.
  • 4Crammer K , Singer Y. On the lesrnability and design of output codes for multiclass problems [A]. Proceedings of the Thirteenth Annual Conference on Computational Learning Theory [ C ]. SanFransisco:Morgan Kanfmann, 2000.35 ~46.
  • 5Hsu C W, Lin C J. A comparison of methods for multiclass support vector machines. hines [ J ]. IEEE Transactions on Neural Networks, 2002,13(2) :415 -425.
  • 6张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):32-42. 被引量:2260
  • 7边肇祺 张学工 等.模式识别[M].北京:清华大学出版社,2001..
  • 8Kreβel U. Pairwise classification and support vector machines [ A]. Advances in Kernel Methods - Support Vector Learning [C]. Cambridge, MA:MIT Press,1999.255 -268.
  • 9刘江华,程君实,陈佳品.支持向量机训练算法综述[J].信息与控制,2002,31(1):45-50. 被引量:96

二级参考文献5

共引文献2390

同被引文献381

引证文献46

二级引证文献288

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