COMPARATIVE ANALYSIS OF METHODS FOR CONSTRUCTING A COMBINED FORECAST
Abstract
This article is devoted to the highly relevant problem of increasing the accuracy of forecasting time series by combining particular forecasts. Forecasting usually uses only one method, while all the information that is contained in other forecasting methods is discarded. Combining forecasts makes it possible to use almost all information contained in particular forecasts. The article describes some most common methods of combining forecasts: two modifications of the Granger-Ramanathan method (without restrictions and with restrictions on the sum of the coefficients for particular forecasts), the method of the pair preference matrix, and the method of linear combination of particular forecasts with different weights.
Authors use the following frequently used time series prediction methods to obtain particular forecasts: the harmonic weights method, the method of adaptive exponential smoothing using the tracking signal, the conventional exponential smoothing method and the Box-Jenkins model.
This analysis is based on production history of several products manufactured in Russia in 1950–2015: the production of electricity; extraction of hard coal; crude oil production; extraction of natural gas; manufacture of metal-cutting machine tools; meat production; production of vegetable oil. The authors carried out comparative analysis of the statistical characteristics of different combined forecasts. Also, the authors calculated the forecast for each of the indicators for 2016 and compared it with the actual data.
The following conclusions were obtained. The combined forecast has a higher accuracy of forecasting time series. Forecasts constructed using the Granger-Ramanathan approaches have the greatest accuracy of the combined forecast.
About the Authors
A. A. Frenkel
Institute of Economics, Russian Academy of Sciences (RAS)
Russian Federation
Moscow
Russian Federation
Moscow
N. N. Volkova
Fund «SONAR»
Russian Federation
Moscow
Russian Federation
Moscow
A. A. Surkov
Fund «SONAR»
Russian Federation
Moscow
Russian Federation
Moscow
E. I. Romanyuk
Institute of Economics, Russian Academy of Sciences (RAS)
Russian Federation
Moscow
Russian Federation
Moscow
References
1. Berezovskaja M., Rajskaja N., Frenkel’ A., Gorjacheva I. Agregirovannyj indeks-jeffektivnyj izmeritel’ infljacii [The aggregated index is an effective measure of inflation]. Voprosy statistiki, 1996, no. 12, pp. 22-25 (In Russ.)
2. Ershov Je.B. [On one method of combining private forecasts]. In: Statisticheskie metody analiza jekonomicheskoj dinamiki. Uch. zap. po statistike, T. XXII-XXIII [Statistical methods of analysis of economic dynamics. Study notes on statistics, vol. XXII-XXIII]. 1973, pp. 87-105 (In Russ.).
3. Surkov A.A. Povyshenie tochnosti jekonomicheskogo prognozirovanija na osnove ob#edinenija jekonomicheskih prognozov [Improving the accuracy of economic forecasting by combining economic forecasts]. Vestnik MFJuA, 2016, no. 2, pp. 267-276 (In Russ.).
4. Frenkel’ A., Rajskaja N., Sergienko Ja., Matveeva O. Indikator jekonomiki [Indicator of economy]. Jekonomicheskie strategii, 2012, no. 9, pp. 32-39 (In Russ.).
5. Frenkel’ A.A. Prognozirovanie proizvoditel’nosti truda: metody i modeli. 2-e izd., dop. i pererab [Forecasting labor productivity: methods and models. Second edition revised and updated]. Moscow, Jekonomika, 2007, 221 p. (In Russ.).
6. Frenkel’ A.A., Volkova N.N., Lobzova A.F., Romanjuk Je.I., Surkov A.A. Ob#edinenie prognozov kak faktor povyshenija kachestva prognozirovanija [Combining forecasts as a factor in improving forecasting quality]. Jekonomika i predprinimatel’stvo, 2016, no. 11, ch.2, pp. 1118-1126. (In Russ.).
7. Bates J.M., Granger C.W.J. The combination of forecasts. Operational Research Quarterly. 1969. Vol. 20. P. 451-468.
8. Clemen R.T. Linear constraints and the efficiency of combined forecasts. Journal of Forecasting. 1986. Vol. 5. P. 31-38.
9. Granger C.W.J., Ramanathan R. Improved methods of combining forecasts. Journal of Forecasting. 1984. Vol. 3. P. 197-204.
10. Gupta S., Wilton P.C. Combination of forecasts: an extension. Management Science. 1987. Vol. 3. P. 356-371.
11. Holden, K., Peel D.A. An empirical investigation of combinations of economic forecasts. Journal of Forecasting. 1986. Vol. 5. P. 229-242.
12. Holden K., Peel D.A., Thomson J.L. Economic Forecasting: An Introduction. New York: Cambridge University Press, 1990. 213 p.
13. Mills T.C., Stepherson M.J. Forecasting contemporaneous aggregates and the combination of forecasts: The case of the U.K. monetary aggregates. Journal of Forecasting. 1985. Vol. 4. P. 273-281.
14. Newbold P., Granger C.W.J. Experience with forecasting univariate time series and the combination of forecasts. J. R. Statist. Soc. 1974. Vol. 137. P. 131-164.
Review
For citations:
Frenkel A.A., Volkova N.N., Surkov A.A., Romanyuk E.I. COMPARATIVE ANALYSIS OF METHODS FOR CONSTRUCTING A COMBINED FORECAST. Voprosy statistiki. 2017;(7):17-27. (In Russ.)
Views:
973
Email this article 