Fast Tuning of SVM Kernel Parameter Using Distance between Two Classes

摘要：

In the construction of support vector machines(SVM)an important step is to select the optimal kernelparameters.This letter proposes using the distancebetween two classes(DBTC)in the feature space tohelp choose kernel parameters.Based on the proposedmethod,the DBTC function is approximatedaccurately with sigmoid function.The computationcomplexity decreases significantly since training SVMand the test with all parameters are avoided.Empirical comparisons demonstrate that the proposedmethod can choose the parameters precisely,and thecomputation time decreases dramatically

作者: Jiancheng Sun

作者单位: School of Electronics,Jiangxi University of Finance and Economics,Nanchang,Jiangxi 330013,China

会议类型: 国际会议

会议名称: 2008 3rd International Conference on Intelligent System and Knowledge Engineering(第三届智能系统与知识工程国际会议)(ISKE 2008)

会议地点: 厦门

会议语种:英文

页码: 108-113

在线出版日期: 2008-11-17（万方平台首次上网日期，不代表论文的发表时间）