Linear random forest to predict energy consumption

Zammarchi, Gianpaolo

2023-01-01

---

Abstract

Forecasting electricity consumption is a relevant task to ensure that the supply of energy fed into the grid always equals the demand. In this study we compare the performance of random forest and linear random forest in the prediction of daily electricity consumption in Italy. We show that both implementations reach a good performance in this task, with the best results obtained by linear random forest in a model including different features such as lags, difference variables and day - month variables.