Data transforms with exponential smoothing methods of forecasting |
| |
| Affiliation: | 1. Institute of Chemical Process Engineering, Boeblinger Str. 78, 70199 Stuttgart, Germany;2. Department of Mathematics, P.O. Box 80010, 3508 TA Utrecht, The Netherlands;1. Graduate School of Convergence Science and Technology, Seoul National University, Republic of Korea;2. School of Electrical Engineering and Computer Science, ASRI, Seoul National University, Republic of Korea |
| |
| Abstract: | In this paper, transforms are used with exponential smoothing, in the quest for better forecasts. Two types of transforms are explored: those which are applied to a time series directly, and those which are applied indirectly to the prediction errors. The various transforms are tested on a large number of time series from the M3 competition, and ANOVA is applied to the results. We find that the non-transformed time series is significantly worse than some transforms on the monthly data, and on a distribution-based performance measure for both annual and quarterly data. |
| |
| Keywords: | State space models Performance measures ANOVA Maximum likelihood AIC |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
