计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (5): 24-26.

• 研究、探讨 • 上一篇    下一篇

基于差分进化算法的支持向量机参数选择

陈 涛,雍龙泉,邓方安,杨 晓   

  1. 陕西理工学院 数学系,陕西 汉中 723000
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-02-11 发布日期:2011-02-11

Parameters selection of support vector machine based on differential evolution

CHEN Tao,YONG Longquan,DENG Fang’an,YANG Xiao   

  1. Department of Mathematics,Shaanxi University of Technology,Hanzhong,Shaanxi 723000,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-02-11 Published:2011-02-11

摘要: 支持向量机参数是影响其性能的重要因素,为了进一步提高支持向量机分类精度和泛化能力,提出了基于差分进化算法的SVM参数选择。以样本误判率最小为优化准则,利用差分进化算法对SVM参数进行优化选择。实验结果表明,利用差分进化算法选择SVM参数,加快了参数搜索的速度,提高了SVM分类精度,该方法具有良好的鲁棒性和较强的全局寻优能力。

关键词: 支持向量机, 差分进化算法, 参数选择

Abstract: Parameters of support vector machine are important factors to affect the performance of SVM.Parameters selection of SVM based on differential evolution is presented to improve the classification accuracy and generalization ability.Optimal rule is the least misjudgment rate of samples,and the parameters of SVM are optimized by using DE.The simulation results show that parameters selection based on DE can accelerate the pace of the search parameters and improve classification accuracy of SVM.It has good robustness and strong global search capability.

Key words: Support Vector Machine(SVM), Differential Evolution(DE), parameters selection