Article ID: | iaor19961478 |
Country: | Belgium |
Volume: | 36 |
Start Page Number: | 27 |
End Page Number: | 42 |
Publication Date: | Sep 1994 |
Journal: | Cahiers du Centre d'tudes de Recherche Oprationnelle |
Authors: | Broze Laurence, Mlard Guy |
Keywords: | ARIMA processes |
Despite being one of the oldest forecasting methods, general exponential smoothing is one of the least used ones. The lack of interest shown for the method which contrasts with the widespread use of the other methods of exponential smoothing can be put down to several reasons: implementing it is fairly intricate, there is no specilised software and there is no survey showing how effective the method is. Some articles have discussed the real reasons why general exponential smoothing is not up to expectation. In this paper, the authors suggest generalising general exponential smoothing and they propose a rigorous methodology allowing a user to estimate the parameters and compute the forecasts.