Linear versus neural network forecasts for European industrial production series

The value of neural network models in forecasting economic time series has been established for North America, but little work has been undertaken for Europe. This paper considers 24 series measuring the annual change in monthly seasonally unadjusted industrial production for important sectors of th...

Full description

Saved in:
Bibliographic Details
Published inInternational journal of forecasting Vol. 20; no. 3; pp. 435 - 446
Main Authors Heravi, Saeed, Osborn, Denise R., Birchenhall, C.R.
Format Journal Article
LanguageEnglish
Published Amsterdam Elsevier B.V 01.07.2004
Elsevier
Elsevier Sequoia S.A
SeriesInternational Journal of Forecasting
Subjects
Economic forecasting
Forecasting
Industrial production
Neural networks
Time series
United Kingdom > UK
France
Germany
Industrial production
Neural networks
Forecasting
Online AccessGet full text

Cover

More Information
Summary:The value of neural network models in forecasting economic time series has been established for North America, but little work has been undertaken for Europe. This paper considers 24 series measuring the annual change in monthly seasonally unadjusted industrial production for important sectors of the German, French and UK economies. Preliminary testing indicates relatively little evidence of nonlinearity in most series. According to root mean-square error (RMSE), linear models generally produce more accurate post-sample forecasts than neural network models at horizons of up to a year. This applies overall and also to the sub-group of series with substantial sample period evidence of nonlinearity. In contrast, the neural network models dominate linear ones in predicting the direction of change. Therefore, the model chosen by users should depend on the type of forecasts they require.
ISSN:0169-2070
1872-8200
DOI:10.1016/S0169-2070(03)00062-1