Electricity Consumption Prediction: Impact of Seasonal Variations with Linear Regression and Neural Networks

Abstract

Meteorological factors, particularly temperature, have a considerable influence on the energy consumption of the Algerian population. Using two real databases, one on temperatures and the other on electricity consumption, covering more than eight consecutive years, the study examines seasonal variations and the impact of temperature on energy demand. Before turning to prediction, a visual analysis is performed to identify trends, using heat maps and time-series graphs. The methodology relies on linear regression and artificial neural network (ANN) models to predict electricity consumption, using temperature and previous day's consumption as the main factors. The results highlight the effects of weather fluctuations on energy demand, providing a basis for energy management strategies and future research into demand forecasting