Application of Box-Jenkins models and artificial neural networks for predicting Gross Domestic Product (GDP) in Algeria

Abstract

This study aims to analyze the economic developments in Algeria’s Gross Domestic Product (GDP) over the period from 1960 to 2023 and to forecast its trajectory through to 2028. To achieve this objective, two different methodologies were employed: the traditional Box-Jenkins approach for time series analysis, and Artificial Neural Networks (ANNs), which are based on artificial intelligence techniques. The aim was to compare their performance in forecasting GDP. The findings of the study revealed that the Box-Jenkins model outperformed the neural networks in terms of forecasting accuracy and quality, as demonstrated by the forecast evaluation metrics and the weighted average used for comparing the two models