LP-SVR Model Selection Using an Inexact Globalized Quasi-Newton Strategy 被引量：1

LP-SVR Model Selection Using an Inexact Globalized Quasi-Newton Strategy

下载PDF

导出

摘要 In this paper we study the problem of model selection for a linear programming-based support vector machine for regression. We propose generalized method that is based on a quasi-Newton method that uses a globalization strategy and an inexact computation of first order information. We explore the case of two-class, multi-class, and regression problems. Simulation results among standard datasets suggest that the algorithm achieves insignificant variability when measuring residual statistical properties. In this paper we study the problem of model selection for a linear programming-based support vector machine for regression. We propose generalized method that is based on a quasi-Newton method that uses a globalization strategy and an inexact computation of first order information. We explore the case of two-class, multi-class, and regression problems. Simulation results among standard datasets suggest that the algorithm achieves insignificant variability when measuring residual statistical properties.

作者 Pablo Rivas-Perea Juan Cota-Ruiz J. A. Perez Venzor David Garcia Chaparro Jose-Gerardo Rosiles

机构地区 Department of Computer Science Department of Electrical and Computer Engineering Science Applications International Corporation

出处《Journal of Intelligent Learning Systems and Applications》 2013年第1期19-28,共10页 智能学习系统与应用（英文）

关键词 Hyper-Parameter Estimation Support VECTOR Regression Machine Learning Data MINING Hyper-Parameter Estimation Support Vector Regression Machine Learning Data Mining

分类号 R73 [医药卫生—肿瘤]

引文网络
相关文献

参考文献1

1周辉仁,郑丕谔.基于GA和Bootstrap的最小二乘支持向量机参数优选[J].系统仿真学报,2008,20(12):3293-3296. 被引量：6

二级参考文献11

1郑小霞,钱锋.基于PCA和最小二乘支持向量机的软测量建模[J].系统仿真学报,2006,18(3):739-741. 被引量：33
2郭辉,刘贺平,王玲.最小二乘支持向量机参数选择方法及其应用研究[J].系统仿真学报,2006,18(7):2033-2036. 被引量：103
3相征,张太镒,孙建成.基于最小二乘支持向量机的非线性系统建模[J].系统仿真学报,2006,18(9):2684-2687. 被引量：27
4赵吉文,刘永斌,孔凡让,陈军宁.核参数遗传选优的SVM在直线电机建模中的应用[J].系统仿真学报,2006,18(12):3547-3549. 被引量：16
5Vapnik V, Levin E, Le Cun Y. Measuring the VC-dimension of learning machines [J]. Neural Computation (S0899-7667), 1994, (6): 851-876.
6Vapnik V N. The nature of statistical learning theory [M]. New York: Springer, 1995.
7J A K Suykens, J Vandewall. Least squares support vector machine classifiers [J]. Neural Processing Letters (S1370-4621), 1999, 9(3): 293-300.
8Pelckmans K, Suykens J A K, De Moor B. Building sparse representations and structure determination on LS-SVM substrates [J]. Neurocomputing (S0925-2312), 2005, 64(S): 137-159.
9G Simon, A Lendasse, V Wertz. Fast approximation of the bootstrap for model selection [C]// ESANN'2003 Proceeding-European Symposium ANN (ISBN2-930307-X). Bruges: D-side public, 2003: 475-480.
10A Lendases, V Wertz, G. Simon. Fast bootstrap applied to LS-SVM for long term prediction of time series [C]// IJCNN'2004 Proceedings-International Joint Conference on NN, Budapest, IEEE. USA: IEEE, 2004: 705-710.