A Comparative Analysis of Ensemble Methods for Short-Term Load Forecasting

Forecasting electricity load is essential to the effective administration and functioning of power systems. This paper investigates the efficacy of various ensemble methods in Short-Term Load Forecasting (STLF) using a real-world dataset. We explore the implementation and comparative performance of four advanced ensemble models: Random Forest (RF) with Extra Trees Regressor (ETR), RF with XGBoost (XGB), XGB with ETR, and a Voting Regressor combining all aforementioned models. Our study focuses on developing these models, preprocessing data, and meticulously tuning parameters to optimize forecasting accuracy and robustness. The models are evaluated using different precision matrices such as R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>, Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE). The Voting Regressor model exhibited the best performance in this work, as indicated by their low RMSE and MAE values and high R2 scores.