Comparison of seasonal estimation methods in multi-item short-term forecasting

Comparison of seasonal estimation methods in multi-item short-term forecasting

0.00 Avg rating0 Votes
Article ID: iaor20003863
Country: Netherlands
Volume: 15
Issue: 4
Start Page Number: 431
End Page Number: 443
Publication Date: Oct 1999
Journal: International Journal of Forecasting
Authors: ,
Keywords: seasonality
Abstract:

This paper addresses the issue of estimating seasonal indices for multi-item, short-term forecasting, based upon both individual time series estimates and groups of similar time series. This development of the joint use of individual and group seasonal estimation is extended in two directions. One class of methods is derived from the procedures developed for combining forecasts. The second employs the general class of Stein Rules to obtain shrinkage estimates of seasonal components. A comparative evaluation has been undertaken of several versions of these methods, based upon a sample of retail sales data. The results favour these newly developed methods and provide some interesting insights for practical implementation.

Reviews

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 Your email address will not be published.