Abstract
In this chapter, we review method selection protocols for three of the commonly used exponential smoothing methods. In addition to protocols which have been previously established, we introduce a new protocol, based on serial variation curves, and a modification of a protocol suggested by Gardner and McKenzie. We also introduce two new decision trees, based on the new protocols, to provide simple ways of choosing between exponential smoothing methods for no trend, damped trend and linear trend. Operational rules are determined for the new rules, determined by detailed experimentation on simulated data. We test the new protocols on real data, and compare the results with established protocols and universal application of smoothing methods. The results show the new approaches to be promising, yielding some improvements in forecasting accuracy. In those cases where no improvement was observed, neither was there any deterioration in forecasting accuracy. This confirms that the new rules introduced in this chapter are robust and worthy of consideration for practical service parts applications.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
A consequence of this choice is that a direct comparison with M3-competition results will not be possible.
- 2.
A drawback of the DW statistic has been taken into account. The Durbin-Watson statistic has a gap between the significant positive autocorrelation, representing the DTM model, and not significant autocorrelation, representing the LGM model. Therefore, if the result belongs to that gap it follows that the DW test is inconclusive. Therefore, an operational rule had to be adopted in order to overcome the problem. There were two possibilities: either to allocate the inconclusive time series to the LGM model or to the DTM model. Having analysed the above issue in both cases during the simulation experiment, it was concluded that the penalty in terms of forecast accuracy is lower, if the inconclusive time series are allocated to the DTM model. Since the LGM model is a special case of the DTM model (for Ï•Â =Â 1), it follows that the LGM could be detected, but not vice versa.
References
Adya M, Collopy F, Armstrong JS, Kennedy M (2001) Automatic identification of time series features for rule-based forecasting. Int J Forecast 17:143–157
Akaike H (1974) A new look at statistical model identification. IEEE Trans Autom Control 19:716–723
Armstrong JS (2001) Extrapolation. In: Armstrong JS (ed) Principles of forecasting: a handbook for researchers and practitioners. Kluwer, Norwell
Assimakopoulos V, Nikolopoulos N (2000) The theta model: a decomposition approach to forecasting. Int J Forecast 16:521–530
Atanackov N (2004) Trend forecasting by constrained optimisation and method selection protocols. PhD thesis, Buckinghamshire Chilterns University College, Brunel University
Billah B, Hyndman RJ, Koehler AB (2005) Empirical information criteria for time series forecasting model selection. J Stat Comput Simul 75:830–840
Box GE, Jenkins GM (1970) Time series analysis, forecasting and control. Holden-Day, San Francisco
Boylan JE, Syntetos AA (2008) Forecasting for inventory management of service parts. In: Kobbacy KAH, Murthy DNP (eds) Complex system maintenance handbook. Springer, London
Chatfield C (1988) Apples, oranges and mean square error. Int J Forecast 4:515–518
Chatfield C (1992) A commentary on error measures. Int J Forecast 8:100–102
Collopy F, Armstrong S (1992) Rule-based forecasting: Development and validation of an expert systems approach to combining time-series extrapolations. Manage Sci 38:1394–1414
Commandeur JJF, Koopman SJ (2007) An introduction to state space time series analysis. Oxford University Press, Oxford
Durbin J, Watson GS (1951) Testing for serial correlation in least squares regression. Biometrika 38:159–177
Fildes R (1992) The evaluation of extrapolative forecasting methods. Int J Forecast 8:81–98
Fildes R, Hibon M, Makridakis S, Meade N (1998) Generalising about univariate forecasting methods. Int J Forecast 14:339–358
Fortuin L (1980) The all-time requirements of spare parts for service after sales–theoretical analysis and practical results. Int J Oper Prod Manage 1:59–69
Gardner ES Jr (1999) Note: rule-based forecasting vs damped trend exponential smoothing. Manage Sci 45:1169–1176
Gardner ES Jr (2006) Exponential smoothing: the state of the art, Part II. Int J Forecast 22:637–666
Gardner ES Jr, McKenzie E (1985) Forecasting trends in time series. Manage Sci 31:1237–1246
Gardner ES Jr, McKenzie E (1988) Model identification in exponential smoothing. J Oper Res Soc 39:863–867
Goodrich RL (1990) Applied statistical forecasting. Business Forecast Systems, Inc, Belmont
Goodrich RL (2001) Commercial software in the M3-competition. Int J Forecast 17:560–565
Hannan EJ, Quinn BG (1979) The determination of the order of an autoregression. J R Stat Soc Ser B 41:190–195
Harrison PJ (1967) Exponential smoothing and short-term sales forecasting. Manage Sci 13:821–842
Harvey AC (1984) A unified view of statistical forecasting procedures. J Forecast 3:245–283
Harvey AC (2006) Forecasting with unobserved components time series models. In: Elliott G, Granger CWJ, Timmermann A (eds) Handbook of economic forecasting, vol 1. Elsevier, Amsterdam
Holt CC (1957) Forecasting seasonals and trends by exponentially weighted moving averages. ONR memorandum, 52. Carnegie Institute of Technology, Pittsburgh, PA
Holt CC (2004a) Forecasting seasonals and trends by exponentially weighted moving averages. Int J Forecast 20:5–10
Holt CC (2004b) Author’s retrospective on ‘Forecasting seasonals and trends by exponentially weighted moving averages. Int J Forecast 20:11–13
Hurvich CM, Tsai CL (1989) Regression and time series model selection in small samples. Biometrika 76:297–307
Hyndman RJ, Billah B (2003) Unmasking the Theta method. Int J Forecast 19:287–290
Hyndman RJ, Koehler AB, Ord JK, Snyder RD (2008) Forecasting with exponential smoothing. the state space approach. Springer, Berlin
Makridakis S, Hibon M (2000) The M3-competition: results, conclusions and practical concerns. Int J Forecast 16:451–476
Makridakis S, Wheelright SC, Hyndman RJ (1998) Forecasting: methods and applications, 3rd edn. Wiley, New York
Makridakis S, Assimakopoulos V, Pagourtzi E, Bougioukos N, Petropoulos F, Nikolopoulos K (2008) PYTHIA: an expert forecasting support system. In: Paper presented at the 28th international symposium on forecasting, Nice, France
Meade N (2000) Evidence for the selection of forecasting methods. J Forecast 19:515–535
Newbold P, Granger CWJ (1974) Experience with forecasting univariate time series and the combination of forecasts. J Roy Stat Soc A, Series A 137:131–165
Pegels CC (1969) Exponential forecasting: some new variations. Manage Sci 12:311–315
Roberts SA (1982) A general class of Holt–Winters type forecasting models. Manage Sci 28:808–820
Sanders N (1997) Measuring forecasts accuracy: some practical suggestions. Prod Invent Manage J (First Quarter), pp 43–46
Schwarz G (1978) Estimating the dimension of a model. Ann Stat 6:461–464
Shah C (1997) Model selection in univariate time series forecasting using discriminant analysis. Int J Forecast 13:489–500
Tashman L (2000) Out-of-sample tests of forecasting accuracy: an analysis and review. Int J Forecast 16:437–450
Tashman LJ, Kruk JM (1996) The use of protocols to select exponential smoothing procedures: a reconsideration of forecasting competitions. Int J Forecast 12:235–253
Taylor JW (2003) Exponential smoothing with a damped multiplicative trend. Int J Forecast 19:715–725
Theil H, Wage S (1964) Some observations on adaptive forecasting. Manage Sci 2:189–206
Vokurka RJ, Flores BE, Pearce SL (1996) Automatic feature identification and graphical support in rule-based forecasting: a comparison. Int J Forecast 12:495–512
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer-Verlag London Limited
About this chapter
Cite this chapter
Atanackov, N.N., Boylan, J.E. (2011). Decision Trees for Forecasting Trended Demand. In: Altay, N., Litteral, L. (eds) Service Parts Management. Springer, London. https://doi.org/10.1007/978-0-85729-039-7_3
Download citation
DOI: https://doi.org/10.1007/978-0-85729-039-7_3
Published:
Publisher Name: Springer, London
Print ISBN: 978-0-85729-038-0
Online ISBN: 978-0-85729-039-7
eBook Packages: EngineeringEngineering (R0)