Aplicação do modelo prophet no contexto da previsão de demanda de energia elétrica

Abstract

The current study evaluates the application of the Prophet model, an open-source time series forecasting library developed by Facebook, in projecting the demand for electrical energy, a very important element for the efficient management of the electrical system. The forecasts must be accurate, given the challenges faced by conventional methods against complex variables such as seasonalities and the dynamics of the sector. The adopted methodology involves collecting and analyzing historical energy consumption data, followed by the construction and adjustment of a predictive model using Prophet to forecast future demand trends, with a subsequent comparison of its forecasts with existing models prepared by EPE to analyze its effectiveness. The results reveal the competence of the proposed model in anticipating demand, surpassing other forecasts in certain periods and standing out for its ability to quickly adjust to new data. The conclusion points to the efficiency of the proposed model as an advanced alternative for energy planning, with potential for continuous improvement and practical application in the sector.