Electricity Consumption Forecasting in Algeria: A Comparison of ARIMA and GM (1,1) Models

Abstract

The pivotal role of electrical energy in propelling a nation's economic development cannot be overstated. A deficiency in the efficient consumption of electricity not only acts as a hindrance to economic growth but also constitutes a potential threat to national security. Recognizing the gravity of this relationship, scholars and policymakers have directed substantial attention towards the exploration of electricity consumption patterns and the development of predictive models. Consequently, a diverse array of models and methodologies has been employed to forecast electrical energy consumption. Within this context, the present research delves into a comparative analysis of two prominent forecasting methods, ARIMA and GM (1,1) grey modeling techniques, with a specific focus on predicting electricity consumption in Algeria. The assessment of these models' performances was conducted using the average percentage of absolute MAPE, employing annual data collected from Algeria spanning the extensive period from 1982 to 2020. The research findings underscore the paramount importance of accurate forecasting in this domain. Notably, the ARIMA model (1,1,0) emerged as the frontrunner, exhibiting superior predictive capabilities when juxtaposed with the GM (1,1) model, as evidenced by the MAPE standard. This nuanced examination contributes to the scholarly discourse on electricity consumption prediction, offering insights that can inform strategic decision-making and policy formulation in the pursuit of sustainable and secure energy practices.