|Table of Contents|

Parameter Optimization of Support Vector Machine for Classification Using Niche Genetic Algorithm

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

Issue:
2009年01期
Page:
16-20
Research Field:
Publishing date:

Info

Title:
Parameter Optimization of Support Vector Machine for Classification Using Niche Genetic Algorithm
Author(s):
ZHU NingFENG Zhi-gangWANG Qi
Department of Automatic Measurement and Control,Harbin Institute of Technology,Harbin 150001,China
Keywords:
parameter optimization niche genetic algorithm support vector machine classification sharing functions
PACS:
TP18
DOI:
-
Abstract:
On the basis of establishing the classification performance evaluation function for support vector machine(SVM),this paper analyzes the influence of SVM parameters on its classification performance,proposes a parameter optimization method of SVM for classification using sharing function based niche genetic algorithm(SNGA).In the SNGA approach,the classification performance evaluation function is used to evaluate the SVM generalization performance,the inverse of the classification performance evaluation function is used as the fittness value.The hamming distance between every two individuals is defined as the sharing function.The method is experimented with five benchmark repositories collected by Gunna Ratsch.The results demonstrate that the algorithm can get the SVM for classification with the best recognition accuracy and simple structure.

References:

[ 1] Vapn ik V. The nature of statistics learn ing theo ry[M ].New York: SpringerVer lag, 1995.

[ 2] Vapn ik V. Statistical lea rning theory [M ]. New York:JohnW iley, 1998.
[ 3] 于德介, 陈森峰, 程军圣, 等. 一种基于经验模式分解与支持向量机的转子故障诊断方法[ J]. 中国电机工程学, 2006, 26( 16): 162- 167.
[ 4] 张晓煜, 普杰信, 黄心汉. 基于2DPCA 和支持向量机的人脸检测研究[ J] . 计算机工程与应用, 2006,2: 94- 197.
[ 5] 张琨, 曹宏鑫, 刘凤玉, 等. 一种基于SVM 的网络入侵检测模型[ J]. 南京理工大学学报(自然科学版),2007, 31( 4): 403- 408.
[ 6] 张琨, 曹宏鑫, 衷宜, 等. 网络入侵检测中一种新型SVM 特征加权分类方法[ J]. 南京理工大学学报( 自然科学版), 2008, 32( 2): 231- 236.
[ 7] W angW en jian, Xu Zongben, LuW e izhen, et a.l Determina tion o f the spread parame ter in theGaussian kernelfo r classification and regression [ J]. Neurocompu-ting, 2003, 55: 653- 663.
[ 8] Stae lin C. Paramete r selection for support vector machines[ DB /OL ]. http: / /www. hp.l hp. com / techreports/2002 /HPL-2002-354R1. htm .l 2008- 01- 06.
[ 9] 郭辉, 刘贺平, 王玲. 最小二乘支持向量机参数选择方法及其应用研究[ J]. 系统仿真学报, 2006, 18( 7): 2033- 2037.
[ 10] 宋乃华, 邢清华. 基于支持向量机的空中目标大类别分类[ J]. 系统工程与电子技术, 2006, 28( 8): 1279- 1281.
[ 11] Zhao Q, Pr incipe J C. Support vector m ach ines fo rSAR autom atic targ et recognition[ J]. IEEE Transactionson Aero space and E lectronic System s, 2001, 37( 2): 643- 654.
[ 12] Broom head D S, Low e D. Mu ltiva riab le functional interpola tion and adaptive netwo rks[ J]. Comp lex Systems, 1988, 2: 321- 355.
[ 13] 朱永生, 张优云. 支持向量机分类器中几个问题的研究[ J]. 计算机工程与应用, 2003, 13: 36- 38.
[ 14] Ratsch G. Benchm ark data sets [ DB /DL] . http: / /ida. first. fraunhofer. de /pro jects/bench/.
[ 15] Raetsch G, Onoda T, Mue ller K-R. Soft marg ins for adaboost[J]. M ach ine Learning, 2001, 42( 3): 287- 320

Memo

Memo:
-
Last Update: 2012-11-19