The Heuristic Algorithms for Selecting the Parameters of Support Vector Machine for Classification  

作　　者：LANG Rongling DENG Xiaole GAO Fei 

机构地区：[1]School of Electronic and Information Engineering, Beihang University, Beijing 100191, China

出　　处：《Chinese Journal of Electronics》2012年第3期485-488,共4页电子学报（英文版）

基　　金：Manuscript Received Apr. 2011; Accepted May 2011. This work is supported by the National Natural Science Foundation of China （No.60702011, No.61071139）, the WeiShi Foundation （YWF-11-03-Q-007） and ＂New Star in Blue Sky＂ Program Foundation.

摘　　要：The performance of Gaussian kernel Sup- port vector machine （SVM） for classification is determined by scale parameter of Gaussian kernel function and error penalty parameter C. A heuristic approach is proposed to tune the parameters of SVM in this paper. We firstly se- lect , and then search the optimal value of C with given . By viewing selection of as a recognition problem, we determine the reasonable range of c using Fisher statistical expression. In selection of C, the search interval is chosen according to the Sequential minimal optimization （SMO） procedure, and the searching procedure is terminated with considering the balance between generalization capability and approximation capability of SVM. The proposed ap- proach is evaluated with a series of real-world data sets.

关 键 词：Support vector machine （SVM） Param-eters selection Classification. 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TN915[自动化与计算机技术—控制科学与工程]