自适应误差惩罚支撑向量回归机

陈晓峰; 王士同; 曹苏群

doi:10.3724/SP.J.1146.2006.01081

自适应误差惩罚支撑向量回归机

doi: 10.3724/SP.J.1146.2006.01081

陈晓峰^{①;王士同},
王士同,
曹苏群

基金项目:

2004年教育部优秀人才支持计划(NCET-04-0496)，模式识别国家重点实验室开放课题，南京大学软件新技术国家重点实验室开放课题，教育部重点科学研究项目(105087)和国防应用基础研究基金项目(A1420061266)资助课题

计量
- 文章访问数: 3138
- HTML全文浏览量: 80
- PDF下载量: 858
- 被引次数: 10
出版历程
- 收稿日期: 2006-07-20
- 修回日期: 2007-01-31
- 刊出日期: 2008-02-19

SVR with Adaptive Error Penalization

摘要

摘要: 该文提出一种支撑向量回归机AEPSVR。它首先用 -SVR求得一个近似的支撑向量回归函数，在此基础上，引入一种新自适应误差惩罚函数，通过迭代，得到鲁棒的支撑向量回归机。该方法因以 -SVR为基础，故可以应用各种求解SVR的优化算法。实验表明，该支撑向量回归机AEPSVR能显著地降低离群点的影响，具有良好的泛化性能。
- 支撑向量回归; 离群点; 自适应误差惩罚
Abstract: A novel support vector regression method AEPSVR is proposed in this paper. First, an approximate regression function is obtained using -SVR method, and then a new adaptive error penalization function is introduced to enhance the robust performance of SVR such that a robust support vector regression is derived. Because the proposed AEPSVR here is based on -SVR, so various optimization methods for SVR can be used. Experimental results show that the proposed AEPSVR can reduce the affect of outliers, and have the very good generalization capability.

HTML全文

参考文献(1)

Smola A J and Schkopf B. A tutorial on support vectorregression[J].Statistics and Computing.2004, 14(3):199-222[2]Schokopf B, Smola A J and Williamson R C, et al.. Newsupport vector algorithm[J].Neural Computation.2000, 12(12):1207-1245[3]Song Q, Hu W, and Xie W. Robust support vector machinewith bullet hole image classification[J].IEEE Trans. on Systems,Man and Cybernetics C.2002, 32(4):440-448[4]Weston J and Herbrich R. Adaptive margin support vectormachines. Neural Information Processing Systems (NIPS)Conference Workshop on Advance in Large MarginClassifiers, Breckenridge Colorado USA, 1998: 281-296.[5]Xu Linli, Crammer K, and Schuurmans D. Robust supportvector machine training via convex outlier ablation. InProceedings of the 21st National Conference on ArtificialIntelligence(AAAI-06), Boston USA, 2006: 536-546.[6]Zhan Yiqiang and Shen Dinggang. An adaptive errorpenalization method for training an efficient and generalizedSVM[J].Pattern Recognition.2006, 39(3):342-350[7]张讲社，郭高. 加权稳健支撑向量回归方法. 计算机学报,2005, 28(7): 1171-1177.Zhang Jiang-she and Guo Gao. Reweighted robust supportvector regression method. Chinese Journal of Computer, 2005,28(7): 1171-1177.[8]Suykens J A K, De Brahanter J, Lukas L, and Vandewalle J.Weighted least squares support vector machine: robustnessand sparse approximation[J].Neurocomputing.2002, 48(1-4):85-105[9]Chuang C C, Su F F, Jeng J T, and Hsiao C C. Robustsupport regression networks for function approximation withoutliers[J].IEEE Trans. on Neural Networks.2002, 13(6):1322-1330[10]Yong Zhan and HaoZhong Cheng. A robust support vectoralgorithm for harmonic and interharmonic analysis of electricpower system[J].Electric Power Systems Research.2005, 73(3):393-400[11]Wang Shitong, Zhu Jiagang and Chung Fu-Lai, et al..Experimental study on parameter choices in norm-r supportvector regression machines with noisy input[J].Soft Computing.2006, 10(3):219-223[12]Wang Shitong, Zhu Jiagang and Chung Fu-Lai, et al..Theoretically optimal parameter choices for support vectorregression machines with noisy input[J].Soft Computing.2005,9(10):732-741

施引文献

期刊类型引用(6)

1.	王昭, 娄昊, 范九伦. 一种基于模糊C均值的图像超像素分割方法. 微电子学与计算机. 2014(10): 18-21+27 . 百度学术
2.	吴涛. 图像阈值化的自适应粗糙熵方法. 中国图象图形学报. 2014(01): 1-10 . 百度学术
3.	罗亮, 冯象初, 张选德, 李小平. 基于非局部双边随机投影低秩逼近图像去噪算法. 电子与信息学报. 2013(01): 99-105 . 本站查看
4.	李伟斌, 高二, 宋松和. 一种全局最小化的图像分割方法. 电子与信息学报. 2013(04): 791-796 . 本站查看
5.	徐东, 彭真明. 基于C-V模型的改进快速水平集图像分割法. 强激光与粒子束. 2012(12): 2817-2821 . 百度学术
6.	董光辉, 席志红, 赵彦青. 基于文化算法的C-V水平集图像分割. 系统工程与电子技术. 2012(07): 1499-1504 . 百度学术