Falak Alsataihi * ⁽¹⁾, Mohamad Taher Anan ⁽²⁾ and Amjad Masso⁽³⁾

(1). Department of Basic Sciences, Faculty of Chemical and Petroleum Engineering, AlBaath University, Homs, Syria.

(2). Department of mathematical Statistics and Programming, Faculty of Science, Aleppo University, Aleppo, Syria.

(*Corresponding author: Falak Alsataihi, E-Mail: falakalsataihi@yahoo.com).

Received: 31/03/2021                     Accepted: 4/07/2021

Abstract

The research studied the methods of generating short-term hierarchical time-series forecasts and applying them to soft wheat production data in Syria, which were classified by region and then by city for time-series from 2006-2018. Where four methods were discussed, namely Top-down, Bottom-up, Middle-out and Optimal Combination, and one of the criteria for testing the quality of prediction methods, which is MASE, was applied to find out the best-performing method to be adopted in research related to this research in order to facilitate and speed up the completion of calculations. The results showed that all four methods differ from each other in general. As a result of the pilot study, it was found that the performance of the Top-down method is better than the performance of the rest of the methods in forecasting, and it was predicted for three future years after 2018, where the results indicated an expectation of an increase in the production of the soft wheat crop at all levels of the hierarchical structure of the studied series.

Keywords: Hierarchical Time Series, Bottom-up approach, Top-down approaches, Middle-out approach, Optimal combination forecasts.

Full paper in Arabic: pdf