文档介绍：第 27 卷第 3 期控制理论与应用 Vol. 27 No. 3
2010 年 3 月 Control Theory & Applications Mar. 2010
文文文章章章编编编号号号: 1000−8152(2010)03−0303−07
基基基于于于混混混合合合最最最小小小二二二乘乘乘支支支持持持向向向量量量机机机网网网络络络模模模型型型的的的非非非线线线性性性系系系统统统辨辨辨识识识
陈杰, 朱琳
(北京理工大学信息科学技术学院自动控制系, 北京 100081)
摘要: 针对基于输入输出数据的非线性系统辨识问题, 提出一种新的混合最小二乘支持向量机(LS-SVMs)网络模
型及相应的学习算法. 该算法将系统的辨识问题动态自适应的划分为若干子问题, 将支持向量机(SVM)用于各子模
块辨识; 通过分析模型的统计学特性, 给出基于整体框架优化的系统参数辨识方法. 针对系统中参数相关联的特性,
采用期望条件最大化(ECM)算法对其进行条件辨识, 同时结合正则化理论和最小二乘法, 保证各专家模块的结构风
险最小化辨识原则. 试验结果表明, 该方法兼具良好的辨识精度和泛化性能.
关键词: 混合专家系统; 最小二乘支持向量机; 非线性系统辨识; 期望条件最大化; 正则化
中图分类号: TP183 文献标识码: A
New identification approach for nonlinear systems based on
work model of
least squares and support vector machines
CHEN Jie, ZHU lin
(Department of Automatic control, School of Information Science and Technology,
Beijing Institute of Technology, Beijing 100081, China)
Abstract: A work model of least squares and support vector machines(MLS-SVMs) and the as-
sociate learning algorithm for identifying nonlinear systems based on the input-output data are proposed. In the model, the
identification task is dynamically posed into several subtasks according to the physical or statistical natures of the
problem. The SVMs are applied as learning machines to every subtask. After analyzing the statistical characteristics of
the model in the formal characterization, we give an algorithm for tr