GENERALIZING ABOUT UNIVARIATE FORECASTING METHODS - FURTHER EMPIRICAL-EVIDENCE

This paper extends the empirical evidence on the forecasting accuracy of extrapolative methods. The robustness of the major conclusions of t he M-Competition data is examined in the context of the telecommunicat ions data of Fildes (1992). The performance of Robust Trend, found to be a successful method for forecasting the telecommunications data by Fildes, is compared with that of other successful methods using the Ra -Competition data. Although it is established that the structure of th e telecommunications data is more homogeneous than that of the M-Compe tition data, the major conclusions of the M-Competition continue to ho ld for this new data set. In addition, while the Robust Trend method i s confirmed to be the best performing method for the telecommunication s data, for the 1001 M-Competition series, this method is outperformed by methods such as Single or Damped Smoothing. The performance of smo othing methods depended on how the smoothing parameters are estimated. Optimisation at each time origin was more accurate than optimisation at the first time origin, which in turn is shown to be superior to arb itrary (literature based) fixed values. In contrast to the last point, a data based choice of fixed smoothing constants from a cross-section al study of the time series was found to perform well. (C) 1998 Elsevi er Science B.V. All rights reserved.