Please wait a minute...
FITEE
Front. Inform. Technol. Electron. Eng.  2011, Vol. 12 Issue (11): 885-896    DOI: 10.1631/jzus.C1100006
    
Novel linear search for support vector machine parameter selection
Hong-xia Pang, Wen-de Dong, Zhi-hai Xu, Hua-jun Feng*, Qi Li, Yue-ting Chen
State Key Laboratory of Optical Instrumentation, Zhejiang University, Hangzhou 310027, China
Novel linear search for support vector machine parameter selection
Hong-xia Pang, Wen-de Dong, Zhi-hai Xu, Hua-jun Feng*, Qi Li, Yue-ting Chen
State Key Laboratory of Optical Instrumentation, Zhejiang University, Hangzhou 310027, China
 全文: PDF(959 KB)  
摘要： Selecting the optimal parameters for support vector machine (SVM) has long been a hot research topic. Aiming for support vector classification/regression (SVC/SVR) with the radial basis function (RBF) kernel, we summarize the rough line rule of the penalty parameter and kernel width, and propose a novel linear search method to obtain these two optimal parameters. We use a direct-setting method with thresholds to set the epsilon parameter of SVR. The proposed method directly locates the right search field, which greatly saves computing time and achieves a stable, high accuracy. The method is more competitive for both SVC and SVR. It is easy to use and feasible for a new data set without any adjustments, since it requires no parameters to set.
关键词： Support vector machine (SVM);  Rough line rule;  Parameter selection;  Linear search;  Motion prediction    
Abstract: Selecting the optimal parameters for support vector machine (SVM) has long been a hot research topic. Aiming for support vector classification/regression (SVC/SVR) with the radial basis function (RBF) kernel, we summarize the rough line rule of the penalty parameter and kernel width, and propose a novel linear search method to obtain these two optimal parameters. We use a direct-setting method with thresholds to set the epsilon parameter of SVR. The proposed method directly locates the right search field, which greatly saves computing time and achieves a stable, high accuracy. The method is more competitive for both SVC and SVR. It is easy to use and feasible for a new data set without any adjustments, since it requires no parameters to set.
Key words: Support vector machine (SVM)    Rough line rule    Parameter selection    Linear search    Motion prediction
收稿日期: 2011-01-05 出版日期: 2011-11-04
CLC:  TP181  
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章　　
Hong-xia Pang
Wen-de Dong
Zhi-hai Xu
Hua-jun Feng
Qi Li
Yue-ting Chen

引用本文:

Hong-xia Pang, Wen-de Dong, Zhi-hai Xu, Hua-jun Feng, Qi Li, Yue-ting Chen. Novel linear search for support vector machine parameter selection. Front. Inform. Technol. Electron. Eng., 2011, 12(11): 885-896.

链接本文:

http://www.zjujournals.com/xueshu/fitee/CN/10.1631/jzus.C1100006    或    http://www.zjujournals.com/xueshu/fitee/CN/Y2011/V12/I11/885

[1] Jian Shi, Shu-you Zhang, Le-miao Qiu. Credit scoring by feature-weighted support vector machines[J]. Front. Inform. Technol. Electron. Eng., 2013, 14(3): 197-204.
[2] Zhi-yong Yan, Cong-fu Xu, Yun-he Pan. Improving naive Bayes classifier by dividing its decision regions[J]. Front. Inform. Technol. Electron. Eng., 2011, 12(8): 647-657.
[3] Wen-de Dong, Yue-ting Chen, Zhi-hai Xu, Hua-jun Feng, Qi Li. Image stabilization with support vector machine[J]. Front. Inform. Technol. Electron. Eng., 2011, 12(6): 478-485.
[4] Wei-dong Chen, Jian-hui Zhang, Ji-cai Zhang, Yi Li, Yu Qi, Yu Su, Bian Wu, Shao-min Zhang, Jian-hua Dai, Xiao-xiang Zheng, Dong-rong Xu. A P300 based online brain-computer interface system for virtual hand control[J]. Front. Inform. Technol. Electron. Eng., 2010, 11(8): 587-597.
[5] Kui-kang Cao, Hai-bin Shen, Hua-feng Chen. A parallel and scalable digital architecture for training support vector machines[J]. Front. Inform. Technol. Electron. Eng., 2010, 11(8): 620-628.
[6] Hyeon Chang Lee, Byung Jun Kang, Eui Chul Lee, Kang Ryoung Park. Finger vein recognition using weighted local binary pattern code based on a support vector machine   [J]. Front. Inform. Technol. Electron. Eng., 2010, 11(7): 514-524.