RNTI

MODULAD
Apprentissage du signal prix de l'électricité. Arbres de régression, séries temporelles et prédictions à long terme
In EGC 2016, vol. RNTI-E-30, pp.543-544
Abstract
Predicting the price of the electricity commodity in the long term is a challenge that current techniques do not meet satisfactorily (Karakatsani et Bunn, 2010; Weron, 2014). In this paper, we introduce a new regression tree based model that yields good predictions on a long-term period with low computational resources requirements. Our approach is validated by temporal series collected from an electricity provider.