[1]郭金玲,王文剑.一种基于数据分布的SVM核选择方法[J].华侨大学学报(自然科学版),2013,34(5):525-528.[doi:10.11830/ISSN.1000-5013.2013.05.0525]
　GUO Jin-ling,WANG Wen-jian.A SVM Kernel Selection Approach Based on the Characteristics of Data Distribution[J].Journal of Huaqiao University(Natural Science),2013,34(5):525-528.[doi:10.11830/ISSN.1000-5013.2013.05.0525]
点击复制

一种基于数据分布的SVM核选择方法()
分享到：

《华侨大学学报（自然科学版）》[ISSN:1000-5013/CN:35-1079/N]

卷:
第34卷
期数:
2013年第5期
页码:
525-528
栏目:
出版日期:
2013-09-20

文章信息/Info

Title:
A SVM Kernel Selection Approach Based on the Characteristics of Data Distribution
文章编号:
1000-5013(2013)05-0525-04
作者:
郭金玲1;  王文剑2; 3
1. 山西大学 商务学院, 山西 太原 030031;2. 山西大学 计算机与信息技术学院, 山西 太原 030006;3. 山西大学 计算智能与中文信息处理教育部重点实验室, 山西 太原 030006
Author(s):
GUO Jin-ling1;  WANG Wen-jian2; 3
1. Business College of Shanxi University, Taiyuan 030031, China; 2. School of Computer and Information Technology, Shanxi University, Taiyuan 030006, China; 3. Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of
关键词:
支撑向量机;  核函数;  核选择;  数据分布;  多维尺度
Keywords:
support vector machine;  kernel function;  kernel selection;  data distribution;  multidimensional scaling
分类号:
TP301
DOI:
10.11830/ISSN.1000-5013.2013.05.0525
文献标志码:
A
摘要:
针对目前支撑向量机(SVM)核函数的选择没有统一规则的现状,提出一种结合数据分布特征进行SVM核选择的方法.首先,采用多维尺度(MDS)分析方法对高维数据集合理降维,提出判断数据集是否呈圆球分布的算法;然后,在得到数据集分布特征的基础上进行SVM核选择,以达到结合数据分布特征合理选择SVM核函数的目的.实验结果表明:呈圆球分布的数据集采用球面坐标核进行分类,识别率达到100%,训练时间最短,优于采用高斯核SVM及多项式核SVM的分类效果.
Abstract:
The kernel selection has no unified rules for support vector machine(SVM). Based on the characteristics of dataset distribution, a new way to select the kernel function was presented. First dimension reduction of the high dimensional dataset was processed with multidimensional scaling(MDS)method. Then an algorithm was put forward, it was judged whether dataset is sphericity distribution. On the basis of determining sphericity distribution, how to select the kernel function was discussed, to achieve the purpose of selecting SVM kernel function with data distribution characteristics. The experimental results illustrate that the classification recognition rate of sphericity datasets reaches 100% with sphere kernel and the training time is the shortest. The classification effect is better than that of using gaussian kernel SVM and polynomial kernel SVM.

参考文献/References:

[1] VAPNIK V.The nature of statitiscal learning theory[M].New York:Spring Verlag Press,1995:4-15.
[2] WANG Wen-jian,XU Zong-ben,LU Wei-zhen,et al.Determination of the spread parameter in the Gaussian kernel for classification and regression[J].Neurocomputing,2003,55(3/4):643-663.
[3] 孙建涛,郭崇慧,陆玉昌,等.多项式核支持向量机文本分类器泛化性能分析[J].计算机研究与发展,2004,41(8):1321-1326.
[4] 张莉,周伟达,焦李成.一类新的支撑矢量机核[J].软件学报,2002,13(4):713-718.
[5] WANG Jin-jun,YANG Jian-chao,YU Kai,et al.Locality constrained linear coding for image classification[J].CVPR,2010,15(3):456-470.
[6] WANG Xiao-ming,CHUNG Fu-lai,WANG Shi-tong.Theoretical analysis for solution of support vector data description[J].Neural Networks,2011,24(4):360-369.
[7] ZAFEIRIOU S,TEFAS A,PITAS I.Minimum class variance support vector machines[J].IEEE Transanctions on Image Processing,2007,16(10):2551-2564.
[8] ZHOU Xi,CUI Na,LI Zhen,et al.Hierarchical gaussianization for image classification[J].ICCV,2009,18(3):79-90.
[9] HUANG Kai-zhu,YANG Hai-qing,KING I,et al.Maxi-min margin machine: Learning large margin classifiers locally and globally[J].IEEE Transanctions on Neural Networks,2008,19(2):260-272.
[10] YU Kai,ZHANG Tong,GONG Yi-hong.Nonlinear learning using local coordinate coding[J].NIPS,2009,26(8):342-356.
[11] CHOI Y S.Least squares one-class support vector machine[J].Pattern Recognition Letters,2009,30(13):1236-1240.
[12] GAO Sheng-hua,TSANG I W H,CHIA L T,et al.Local features are not lonely laplacian sparse coding for image classification[J].CVPR,2010,18(6):126-138.
[13] 周伟达,张莉,焦李成.一种改进的推广能力度量标准[J].计算机学报,2003,26(5):598-604.
[14] WU Si,AMARI S I.Conformal transformation of kernel functions: A data-dependent way to improve support vector machine classifiers[J].Neural Processing Letters,2002,15(1):59-67.
[15] CHAPELLE O,VAPNIK V.Model selection for support vector machines[C]//Advances in Neural Information Processing Systems 12.Cambridge:MIT Press,2001:120-155.
[16] VAPNIK V.The nature of statitiscal learning theory[M].New York:Spring-Verlag Press,1995:88-125.
[17] COX T,COX M.Multidimensional Scaling[M].London:Chapman & Hall,1994:145-150.

备注/Memo

备注/Memo:
收稿日期: 2013-03-01
通信作者: 郭金玲(1982-),女,讲师,主要从事机器学习与数据挖掘的研究.E-mail:tygjl@163.com.
基金项目: 国家自然科学基金资助项目(61273291); 山西省高等学校科技研究开发项目(20121131); 山西大学商务学院科研基金资助项目(2012013)
更新日期/Last Update: 2013-09-20