Time series forecasting using neural networks: Should the data be deseasonalized first?

Time series forecasting using neural networks: Should the data be deseasonalized first?

0.00 Avg rating0 Votes
Article ID: iaor20012571
Country: United Kingdom
Volume: 18
Issue: 5
Start Page Number: 359
End Page Number: 367
Publication Date: Sep 1999
Journal: International Journal of Forecasting
Authors: , , ,
Keywords: neural networks
Abstract:

This research investigates whether prior statistical deseasonalization of data is necessary to produce more accurate neural network forecasts. Neural networks trained with deseasonalized data from Hill et al. were compared with neural networks estimated without prior deseasonalization. Both sets of neural networks produced forecasts for the 68 monthly time series from the M-competition. Results indicate that when there was seasonality in the time series, forecasts from neural networks estimated on deseasonalized data were significantly more accurate than the forecasts produced by neural networks that were estimated using data which were not deseasonalized. The mixed results from past sudies may be due to inconsistent handling of seasonality. Our findings give guidance to both practitioners and researchers developing neural networks.

Reviews

Required fields are marked *. Your email address will not be published.