METHODOLOGICAL APPROACHES TO IMPROVEMENT OF FORECAST ACCURACY BY COMBINING FORECASTS

The authors introduce problems of improvement of forecast accuracy and offer several ways of their solution, basing on a retrospective analysis of domestic and foreign studies on methodology of socio-economic forecast. In their opinion, a real modern solution to the problem is in implementation of the approach, linked to forecast combination, because it is difficult to prefer one forecasting method to another. Special attention is driven to private forecasts combination technique, to the efficiency of these combination methods and to the optimization of the number of selected forecast options for combination. Combination of forecasts has already proved itself in practice and it is not inferior to the private methods of forecasting in accuracy. The main idea of combining forecasts is the use of all available information regarding various forecasting methods, even if these methods are not sufficiently accurate. This paper is a review of different ways of constructing weights for combining forecasts. Methods of combining forecasts are described as follows: 1) by averaging private forecasts, 2) along with using the method of least squares, 3) involving minimization of an error variance of combined forecast, 4) based on retrospective forecasts 5) based on factor analysis, 6) with a use of paired comparisons, 7) based on sequential quadratic programming. Moreover, advantages and disadvantages of different methods of weighing rates are introduced. The conclusion of the article contains tables with basic and most frequently used methods of combining forecasts and a description of results obtained from these method. There is also a vast bibliography of scientific publications, of both domestic and foreign authors on the subject.

прогнозирование, построение модели прогноза, комбинированный прогноз, объединение прогнозов, весовые коэффициенты, prediction, forecast, combined forecast, the association forecasts, weighting rates

1. Балтрушевич Т.Г. Модели и методы оценки эффективности гибких производственных систем //Автореф. дис..канд. экон. наук. М., 1991. С. 17 - 20.

2. Бейлинсон Я.Е., Мотова М.А. Комбинированные модели прогноза // Экспресс-информация. Серия: Модели рование социально-экономических процессов. М., 1990. Вып. 2. С. 110-121.

3. Васильев А.А. Гибридные модели прогноза экономических показателей на основе взвешенного арифметического среднего простого набора прогнозов //Вестник ТвГУ. Серия «Экономика и управление". 2012. № 13. С. 149 - 165.

4. Васильев А.А. Объединение прогнозов на основе усеченных и винзорированных средних //Вестник ТвГУ. Серия «Экономика и управление». 2014. № 2. С. 204 - 215.

5. Васильев А.А. Объединение прогнозов на основе оценки Ходжеса-Лемана и ее модификациях //Вестник ТвГУ. Серия «Экономика и управление». 2014. № 4. С. 201- 215.

6. Васильев А.А. Генезис гибридных моделей прогнозирования на основе объединения прогнозов //Вестник ТвГУ. Серия «Экономика и управление». 2014. № 23. С. 316 - 331.

7. Головченко В. Б., Носков С. И. Комбинирование прогнозов с учетом экспертной информации //Автоматика и телемеханика. 1992. № 11. С. 109 - 117.

8. Горелик Н.А., Френкель А.А. Статистические проблемы экономического прогнозирования //В кн.: Статистические методы анализа экономической динамики. Уч. зап. по статистике.М.: Наука, 1983. Т. 46. С. 9-48.

9. Дуброва Т.А Статистический анализ и прогнозирование экономической динамики: проблемы и подходы //В кн.: Методология статистического исследования социально-экономических процессов. М.: Юнити, 2012. С. 129-138.

10. Ершов Э.Б. Об одном методе объединения частных прогнозов //В кн.: Статистические методы анализа экономической динамики. Уч. зап. по статистике. М.: Наука, 1973. Т. XXII-XXIII. С. 87-105.

11. Ицхоки О. Выбор модели и парадоксы прогнозирования //Квантиль. 2006. № 1. С. 43-51.

12. Лукашин Ю.П. Адаптивные методы краткосрочного прогнозирования временных рядов. М.: Финансы и статистика, 2003. С. 121-135.

13. Остапюк С.Ф., Мотова М.А Модели построения комбинированного прогноза развития научно-технической сферы // Проблемы прогнозирования. 2004. № 1. С. 146 - 156.

14. Френкель АА Прогнозирование производительности труда: методы и модели. М.: Экономика, 1989. С. 142-154.

15. Френкель А.А., Сурков А.А. Объединение прогнозов - эффективный инструмент повышения точности прогнозирования //Экономист. 2015. № 1. С. 44 - 56.

16. Френкель А.А., Райская H. H., Бурцева Т.А., Сурков А.А. Повышение точности прогнозирования инвестиций на основе объединения различных прогнозов //Научный вестник ИЭП им. Гайдара. 2015. № 15/1. С. 48 -50.

17. Яковенко А.А. Комбинирование результатов прогнозирования временных рядов как метод повышения точности прогноза // Сборник научных трудов НГТУ. 2008. № 2 (52). С. 71-78.

18. Armstrong J.S. Combining forecasts: the end of the beginning or the beginning ofthe end? //International Journal of Forecasting. 1989. Vol. 5. P. 585-588.

19. Armstrong J.S. Combining forecasts // Kluwer Academic Publishers. 2001. P. 1-19.

20. Barnard G.A New methods of quality control //J. R. Statist. Soc. 1963. Vol. 126. P 255-259.

21. Bates J.M. and Granger C.W.J. The combination of forecasts // Operational Research Quarterly. 1969. Vol. 20. P. 451-468.

22. Bessler D.A., Brandt J.A. Forecasting livestock prices with individual and composite methods //Applied Economics. 1981. Vol. 13. P. 513-522.

23. Bunn D.W. A Bayesian approachto the linear combination of forecasts //Operational Research Quarterly. 1975. Vol. 26. P. 325-329.

24. Bunn D.W.A Comparative evaluation of the outperformance and minimum variance procedures for the linear synthesis of forecasts // Operational research quarterly. 1977. Vol. 28. № 3. P. 653-662.

25. Clemen R.T. Linear constraints and the efficiency of combined forecasts // Journal of Forecasting. 1986. Vol. 5. P. 31-38.

26. Clemen R.T. Combining forecasts: a review and annotated bibliography //International Journal of Forecasting. 1989. Vol. 5. P. 559-583.

27. Clemen R.T. and Winkler R.L. Combining economic forecasts // Journal of Business & Economic Statistics. 1986. Vol. 1. P. 39-46.

28. Clements M. P. and Hendry D. F. Explaining the results of the M3 forecasting competition, 2001.

29. Crane D. B. and Crotty J.R. A two-stay forecasting model: Exponential smoothing and multiple regression // Management Science. 1967. Vol. 13. P. 501-507.

30. Croushore D. Introducing the survey of professional forecasters. Business review, Federal Reserve Bank of Philadelphia, 1993.

31. Fang Y. Forecasting combination and encompassing tests // International Journal of Forecasting. 2003. Vol. 19. P 87-94.

32. Fiordaliso A. Combining forecasts a fuzzy approach. International Symposium on Intelligent Systems, 1997.

33. Flores B. E. and White E. M. Subjective versus Objective Combining of Forecasts: an Experiment //Journal of Forecasting. 1989. Vol. 8. P. 331-341.

34. Goodwin P. New evidence on the value of combining forecasts // FORESIGHT. 2009. Vol. 12. P. 33-35.

35. Granger C. W. J. Aggregation of time series variables - a survey // Institute for Empirical Macroeconomics and University of Minnesota, 1988.

36. Granger C. W. J. Invited review: combining forecasts -twenty years later // Journal of Forecasting. 1989. Vol. 8. P. 167-173.

37. Granger C. W. J. and Ramanathan R. Improved methods of combining forecasts // Journal of Forecasting. 1984. Vol. 3. P. 197-204.

38. Gupta S. and Wilton P. C. Combination of forecasts: an extension //Management Science. 1987. Vol. 3. P 356-371.

39. Gupta S. and Wilton P.C. Combination of Economic Forecasts: An Odds-Matrix Approach //Journal of Business and Economic Statistics. 1988. Vol. 6. P 373-379.

40. Hibon M., EvgeniouT. To combine or not to combine: selecting among forecasts and their combinations // International Journal ofForecasting. 2005.Vol. 21. P. 15- 24.

41. Jose V. R. R. and Winkler R. L. Simple robust averages of forecasts: some empirical results //International Journal of Forecasting. 2008. Vol. 24. P. 163-169.

42. Kang H. Unstable weights in the combination of forecasts //Management Science. 1986. Vfol.6. P. 687-695.

43. Mahmoud A. Accuracy in Forecasting: a Surver //Journal ofForecasting. 1984. Vol. 5. P. 139-159.

44. Makridakis S. Why combining works? // International Journal ofForecasting. 1989. Vl. 5. P. 601-603.

45. Makridakis S. and Hibon M. The M3-Competition: results, conclusions and implications // International Journal of Forecasting. 2000. Vl. 16. P. 451-476.

46. Makridakis S. and Winkler R.L. Averages of forecasts: some empirical results //Management Science. 1983. Vol. 9. P. 987-996.

47. Makridakis S., Chatfield C., Hibon M., Lanrence M., Mills T., Ord K., Simmons L.F. The M2-Competition: a real-time judgmentally based forecasting study // International Journal ofForecasting. 1993. Vfol. 16. P. 451-476.

48. Makridakis S. et al. The Accuracy ofExtrapolation (Time Series) Methods: Result ofa Forecasting Competition // Journal Forecasting. 1982. Vfol. 1.P 111-153.

49. Makridakis S. et al. The Accuracy of Major of Extrapolation (Time Series) Methods // Wiley. London (in press). 1983.

50. Mancuso A. C. B. and Werner L.Review of combining forecasts approaches // Independent Journal ofManagement & Production. 2013. Vol. 4. P. 248-277.

51. McIntoch C. S. and Bessler D. A. Forecasting agricultural prices using a Bayesian composite approach // Southern Journal of Agricultural Economics. 1988. Vol. 1. P 73 - 80.

52. Menezes L. M., Bunn D. W. and Taylor J.W. Review of guidelines for the use ofcombined forecasts //European Journal of Operational Research. 2000. Vl.120. P 190-204.

53. Newbold P. and 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.

54. Russell T. D. and Adam T. D. Jr. An empirical evaluation ofalternative forecasting combinations //Management Science. 1987. Vol.10. P. 1267-1276.

55. Smith D. G. C. Combination of forecasts in electricity demand prediction // Journal of Forecasting. 1989. Vol. 8. P. 349-356.

56. Stock J. H. and Watson M. W. Combination forecasts of output growth in a seven-country data set //Journal ofForecasting. 2004. Vfol. 23. P. 405-430.

57. Taylor J. W. and Majithia S. Using combined forecasts with changing weights for electricity demand profiling // Journal ofthe Operational Research Society. 2000. Vl. 51. P 72-82.

58. Trenkler G. and Gotu B. Combination of forecasts: a bibliography // Department of Statistics University of Dortmund, 1998.

59. Trenkler G. and Liski E.P. Linear constraints and the efficiency ofcombined forecasts //Journal ofForecasting. 1986. Vol. 5. P. 197-202.

60. Wallis K. F. Combining forecasts - forty years later // Applied Financial Economics. 2011. Vfol. 21. P. 33-41.

61. Winkler, R. L., &Clemen, R. T. Sensitivity of weights in combining forecasts // Operations Research. 1992. Vfol. 40. P. 609-614.

62. Winkler R. L. and Makridakis S. The combination of forecasts //J. R. Statist. Soc.1983. Vfol. 146. P. 150-157.

63. Yang Y. Combining forecasting procedures: some theoretical results // Department ofStatistics and Statistical Laboratory Iowa State University. 2001a.

64. Yang Y. Adaptive regression by mixing //Journal of American Statistical Association. 2001b. Vol. 96. P 574- 588.

65. Zarnowitz V. An appraisal of short-term economic forecasts // National Bureau ofEconomic Research. New York. 1967. P. 123 - 125.

66. Zou H. and Yang Y. Combining time series models for forecasting // International Journal ofForecasting. 2004. Vol. 20. P. 69-84.