استخدام نماذج الشبكات العصبية والانحدار الذاتي للتنبؤ بالناتج المحلي والصادرات الصناعية السورية

Abstract

This research aims to analyze the performance of the Syrian industrial sector during the period (2000-2021), with a focus on industrial GDP and industrial exports. The study adopted a methodology that combines Autoregressive (AR) models and Artificial Neural Networks (ANN) to handle the volatile nature of the data and build accurate forecasts for the studied industrial indicators up to the year 2027.

The results revealed a sharp decline in the industrial sector's contribution to the GDP from 30.12% to 9.52%, while the share of industrial exports dropped from 17.23% to 3.61% during the study period. The forecasts indicated a modest growth for the industrial GDP, reaching 65,825.01 million SYP in 2024 with an annual growth rate of 2-3%, and a predicted relative stability until 2027 at 65,430.58 million SYP. In contrast, export forecasts suggest a continued decline, albeit at a less volatile pace.

The study demonstrates the efficiency of hybrid models in analyzing unstable economic data and dealing with its sharp fluctuations. It provides a useful analytical framework for policymakers to develop strategies for revitalizing the industrial sector in Syria.