Approximation to Nonlinear Discrete-Time Systems by Recurrent Neural Networks

被引：0
|
作者：
Li, Fengjun [1 ]
机构：
[1] Ningxia Univ, Sch Math & Comp Sci, Yinchuan 750021, Peoples R China
来源：
SIXTH INTERNATIONAL SYMPOSIUM ON NEURAL NETWORKS (ISNN 2009) | 2009年 / 56卷
关键词：
Approximation; nonlinear discrete-time system; recurrent neural networks; UNIVERSAL APPROXIMATORS; DYNAMICAL-SYSTEMS;
D O I：
暂无
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Neural networks are widely used to approximate nonlinear functions. In order to study its approximation capability, a approximating approach for nonlinear discrete-time systems is presented by using the concept of the time-variant recurrent neural networks (RNNs) and the theory of two-dimensional systems. Both theory and simulations results show that the derived mathematical model of RNNs can approximate the nonlinear dynamical systems to any degree of accuracy.
引用
收藏
页码：527 / 534
页数：8
相关论文
共 50 条
12345