| Study on Support Vector Machine Based on 1-Norm |
| |
| 作者姓名： | 潘美芹 贺国平 韩丛英 薛欣 史有群 |
| |
| 作者单位： | [1]College of Information Science and Engineering, Shandong University of Science and Technology, Qingdao 266510 [2]College of Computer Science ＆ Engineering, Donghua University, Shanghai 200051 [3]Department of Mathematics, Taishan University, Tai＇an 271000 |
| |
| 摘 要： | The model of optimization problem for Support Vector Machine（SVM） is provided, which based on the definitions of the dual norm and the distance between a point and its projection onto a given plane. The model of improved Support Vector Machine based on 1-norm （1 - SVM） is provided from the optimization problem, yet it is a discrete programming. With the smoothing technique and optimality knowledge, the discrete programming is changed into a continuous programming. Experimental results show that the algorithm is easy to implement and this method can select and suppress the problem features more efficiently.Illustrative examples show that the 1 - SVM deal with the linear or nonlinear classification well.
|
| 关 键 词： | 支持向量机 特征选择 特征抑制 分析方法 |
| 收稿时间： | 2006-08-20 |
Study on Support Vector Machine Based on 1-Norm |
| |
| PAN Mei-qin,HE Guo-ping,HAN Cong-ying,XUE Xin,SHI You-qun.Study on Support Vector Machine Based on 1-Norm[J].Journal of Donghua University,2006,23(6):148-152. |
| |
| Authors: | PAN Mei-qin HE Guo-ping HAN Cong-ying XUE Xin SHI You-qun |
| |
| Abstract: | The model of optimization problem for Support Vector Machine(SVM) is provided, which based on the definitions of the dual norm and the distance between a point and its projection onto a given plane. The model of improved Support Vector Machine based on 1-norm(1-SVM) is provided from the optimization problem, yet it is a discrete programming. With the smoothing technique and optimality knowledge, the discrete programming is changed into a continuous programming. Experimental results show that the algorithm is easy to implement and this method can select and suppress the problem features more efficiently. Illustrative examples show that the 1-SVM deal with the linear or nonlinear classification well. |
| |
| Keywords: | 1- SVM best separating plane feature suppression feature selection |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
