Two-Sided Exponential Smoothing Method for Restoring of Dynamic Processes

Олена Вікторівна Братусь, Володимир Миколайович Подладчіков, Петро Іванович Бідюк

Abstract


Background. Restoring of the true regularities and missed values of time series is an important stage of data preparation for the future modeling and forecasting, therefore development of new methods of restoring is needed.

Objective. To develop two-sided exponential smoothing method for restoring of regularities of dynamic processes evolution; to apply created method for restoring of missed values of London metal exchange average day prices for color metal (zinc) and to compare with methods of restoring by using arithmetic mean, autoregressive approach and exponential smoothing method.

Methods. To achieve the formulated goal the following methods were used: two-sided exponential smoothing method was created; restoring by using of arithmetic mean values with usage of known values; autoregressive approach and exponential smoothing.

Results. Two-sided exponential smoothing method was developed, which contains procedure of smoothing in direct and reversed time. The proposed method was used for restoring of dynamic processes and missed values of time series. Restoring of missed values of average daily prices for color metal (zinc) by making use of developed method and comparison with other methods were performed.

Conclusions. It is shown by means of simulation that two-sided exponential smoothing method is effective for restoring of process regularities. Developed method for restoring missing values of zinc prices in its application on practice showed an advantage in comparison with all the methods used in this study by the values of statistical characteristics of adequacy for constructed models, so it could be used in practice.

Keywords


Restoring of dynamic processes regularities; Restoring of missed values of time series; Two-sided exponential smoothing; Exponential smoothing; Arithmetic mean; Autoregressive approach

References


R.S. Tsay, Analysis of Financial Time Series. New York: John Wiley & Sons, Inc., 2010.

F.-B. Behrouz, Adaptive Filters. New York: John Wiley & Sons, Inc., 2013.

R.G. Brown, Exponential Smoothing for Predicting Demand. Cambridge, Massachusetts: Arthur D. Little Inc., 1956.

C.C. Holt, “Forecasting trends and seasonal by exponentially weighted averages”, Int. J. Forecasting, vol. 20, no. 1, pp. 5–10, 1957.

E.S. Gardner, Jr., “Exponential smoothing: the state of the art”, J. Forecasting, vol. 4, no. 1, pp. 1–28, 1985.

E.S. Gardner Jr., “Exponential smoothing: The state of the art – Part II”, J. Forecasting, vol. 22, no. 4, pp. 637–666, 2006.

R.J. Hyndman et al., “Prediction intervals for exponential smoothing using two new classes of state space models”, J. Forecasting, vol. 24, no. 1, pp. 17–37, 2005.

S. Kolassa, “Combining exponential smoothing forecasts using Akaike weights”, Int. J. Forecasting, vol. 27, no. 2, pp. 238–251, 2011.

N.D. Pankratova et al., “Quasioptimal smoothing as a tool for the analysis of complex semistructured dynamic processes”, Cybernetics and Systems Analysis, vol. 45, no. 6, pp. 916–923, 2009.

M.Z. Zgurovskyi and V.N. Podladchikov, Analytical Methods of Kalman Filtration for Systems with a Priori Uncertainty. Kyiv, Ukraine: Naukova Dumka, 1995 (in Russian).

Official site of London Metal Exchange. Section “Historical data” [Online]. Available: https://www.lme.com/en-gb/pricing-and-data/ historical-data/

P.I. Biduyk et al., Analysis of Time Series. Kyiv, Ukraine: NTUU KPI, 2013 (in Ukrainian).


GOST Style Citations


  1. Tsay R.S. Analysis of Financial Time Series. – New York: John Wiley & Sons, Inc., 2010. – 715 p.

  2. Behrouz F.-B. Adaptive Filters. – New York: John Wiley & Sons, Inc., 2013. – 802 p.

  3. Brown R.G. Exponential Smoothing for Predicting Demand. –Cambridge, Massachusetts: Arthur D. Little Inc., 1956. – 15 p.

  4. Holt C.C. Forecasting trends and seasonal by exponentially weighted averages // Int. J. Forecasting. – 1957. – 20, № 1. – P. 5–10.

  5. Gardner E.S., Jr. Exponential smoothing: the state of the art // J. Forecasting. – 1985. – 4, № 1. – P.1–28.

  6. Gardner E.S., Jr. Exponential smoothing: the state of the art – Part II // J. Forecasting. – 2006. – 22, № 4. – P. 637–666.

  7. Prediction intervals for exponential smoothing using two new classes of state space models / R.J. Hyndman, A.B. Koehler, J.K. Ord, R.D. Snyder // J. Forecasting. – 2005. – 24, № 1. – P. 17–37.

  8. Kolassa S. Combining exponential smoothing forecasts using Akaike weights // Int. J. Forecasting. – 2011. – 27, № 2. –    P. 238–251.

  9. Pankratova N.D., Podladchikova T.V., Strelkov D.G. Quasioptimal smoothing as a tool for the analysis of complex semistructured dynamic processes // Cybernetics and Systems Analysis. – 2009. – 45, № 6. – P. 916–923.

  10. Згуровский М.З., Подладчиков В.Н. Аналитические методы калмановской фильтрации для систем с априорной неопределенностью. – К.: Наук. думка, 1995. – 283 с.

  11. Офіційний сайт Лондонської біржі металів. Розділ “сторичні дані” [Електронний ресурс]. – https://www.lme.com/en-gb/ pricing-and-data/historical-data/

  12. Бідюк П.І., Романенко В.Д., Тимощук О.Л. Аналіз часових рядів. – К.: НТУУ “КПІ”, 2013. – 600 с.




DOI: https://doi.org/10.20535/1810-0546.2016.6.80305

Refbacks

  • There are currently no refbacks.




Copyright (c) 2017 NTUU KPI