Time series modelling using neural networks with autoregressive-moving average structure

被引：0

作者：

Chan, CW ^{[1
]}

Chan, WC ^{[1
]}

Cheung, KC ^{[1
]}

机构：

[1] Univ Hong Kong, Dept Mech Engn, Hong Kong, Hong Kong

来源：

ALGORITHMS AND ARCHITECTURES FOR REAL-TIME CONTROL 1997 | 1997年

关键词：

neural networks; time series analysis; autoregressive-moving average model;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Though autoregressive-moving average (ARMA) models are widely used in time series analysis, it is not suitable for modelling nonlinear systems. As neural networks has the attractive property of approximating linear or nonlinear functions with arbitrary accuracy, it is a good alternative to the ARMA models. In earlier works, neural networks with AR structure are proposed. In this paper, neural networks with ARMA structure is proposed, and an iterative procedure is devised to train the networks.

引用

页码：87 / 88

页数：2