基于支持向量回归的非线性系统辨识

基金项目: 国家自然科学基金资助项目（60024301）；863计划资助项目（2001AA114202）

详细信息

APPLICATION OF SUPPORT VECTOR REGRESSION TO NONLINEAR SYSTEM IDENTIFICATION

Institute of Artificial Intelligence and Robotics, Xi'an Jiaotong University, Xi'an 710049, China

摘要: 本文将支持向量回归方法应用于非线性系统辨识问题．基于高斯支持向量回归及不敏感损失函数的基本思想，本文提出一个非线性系统辨识的新算法，并将其与用于系统辨识的径向基函数神经网络进行了比较．模拟实验表明，支持向量回归方法可以成为非线性系统辨识的有力工具．
- 支持向量回归 /
- 非线性系统辨识 /
- 贝叶斯证据框架
Abstract: This paper applies Support Vector Regression(SVR) to nonlinear system identification problem. Using the basic idea of Gaussian SVR and -insensitive loss function, we propose a new algorithm for nonlinear system identification and compare the Gaussian SVR with the radial basis function(RBF) network for system identification. The performance of the SVR is illustrated by a simulation example involving a benchmark nonlinear system.
- support vector regression /
- nonlinear system identification /
- Bayesian evidence framework

[1]	Narendra K S, Parthasarathy K. Identification and control of dynamical systems using neural networks [J].IEEE Trans. on Neural Networks, 1990,1:4～27.
[2]	Chan W C, Chan C W, Cheung, K C, et al. On the modelling of nonlinear dynamic systems using support vector neural networks [J].Engineering Applications of Artificial Intelligence, 2001, 14:105～113.
[3]	Vapnik V N. The Nature of Statistical Learning Theory [M].NewYork:Springer-Verlag, 1995.
[4]	Smola A. Learning with kernels [D].Berlin, Germany:Technischen Universit¨at Berlin, 1998.
[5]	Gao J B, Gunn S R, Harris C J. A probabilistic framework for SVM regression and error estimation [J].Machine Learning, 2002,46:71～89.
[6]	Kwok J T-Y. The evidence framework applied to support vector machines [J].IEEE Trans. on Neural Networks, 2000,11:1162～1173.