期刊文献+

核向量机算法研究及应用 被引量:2

Core Vector Machine Algorithm and its Application
下载PDF
导出
摘要 对训练样本规模为m的标准支持向量机(Support Vector Machine,SVM)进行训练,时间复杂度为O(m3),空间复杂度为O(m2)。文章研究将其转换成等价的最小包含球(Minimum Enclosing Ball,MEB)形式,使用核心集向量机(Core Vector Machine,CVM)高效获得近似最优解。CVM的优点是时间复杂度与训练样本规模m呈线性关系,空间复杂度与m无关。实验证明,CVM可以对大规模数据集进行高效的分类。 According to the training set size m,standard SVM training has O(m3) time and O(m2) space complexities.In this paper,CVM algorithm is discussed.It transforms the kernel method into equivalent MEB problems,and gets the approximately optimal solutions efficiently by core set.CVM has a time complexity that is liner in m and a space complexity that is independent of m.The result shows that CVM algorithm can handle much larger data sets than existing scale up methods.
作者 许敏
出处 《无锡职业技术学院学报》 2012年第4期73-76,共4页 Journal of Wuxi Institute of Technology
关键词 核向量机 支持向量机 最小包含球 核函数 Core Vector Machine(CVM) Support Vector Machine(SVM) Minimum Enclosing Ball(MEB) kernel function
  • 相关文献

参考文献3

  • 1Tsang I, Kwok J, Cheung P. Core vector machines: Fast SVM training on very large data sets[J]. J of Machine Learning Research, 2005, 6(4): 363-392.
  • 2Badoiu M, Clarkson K L. Optimal core sets for balls [J]. Computational Geometry: Theory and Applica- tions, 2008,40(1); 14 -22.
  • 3Smola A, Sch61kop{ B. Sparse greedy matrix approxi- mation for machine learning [C]. Stanford, CA: Proc. 7thInt. Conf. Mach. Learn., 2000:911-918.

同被引文献17

引证文献2

二级引证文献4

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