Short-Term Electricity Load Forecasting with Machine Learning

An accurate short-term load forecasting (STLF) is one of the most critical inputs for power plant units’ planning commitment. STLF reduces overall uncertainty added by intermittent production renewable sources; thus, it helps to minimize hydrothermal electricity costs in a grid. Although there some research field and even several applications, continual need improve forecasts. This proposes set machine learning (ML) models accuracy 168 h The developed employ features from multiple sources, such as historical load, weather, holidays. Of five ML tested various profile contexts, Extreme Gradient Boosting Regressor (XGBoost) algorithm showed best results, surpassing previous weekly predictions based on neural networks. Additionally, because XGBoost are an ensemble decision trees, facilitated model’s interpretation, which provided relevant additional result, features’ importance forecasting.