Prediction of wind power generation with XGBoost: case study in Norway
Nordic Machine Intelligence · 12.06.2026 · 6 (2) · 42-47
DOI: 10.5617/nmi.12933 ↗Kurz gefasst
Fachartikel in der Zeitschrift Nordic Machine Intelligence, online erschienen am 12.06.2026. Verfasst von Riccardo Parviero und 4 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Wind power generation is inherently volatile due to its dependence on weather conditions, making accurate forecasting essential for modern energy systems. This study investigates a forecasting approach based on XGBoost, incorporating tailored feature engineering of weather-related inputs to enhance model performance. Furthermore, the use of XGBoost routines permits easy switching between deterministic and probabilistic forecasting, the latter achieved by applying quantile regression. The proposed forecasting pipeline is tested for forecasting wind power generation across Norwegian bidding zones, yielding better results on both Mean Absolute Error and Pinball loss against a pre-existing baseline, evaluated on walk-forward hold-out data spanning all four quarters of the latest year available in the dataset.
Zitieren
APA
Parviero, R., He, X., Hashemi, N., Topka, W., & Martinelli, G. (2026). Prediction of wind power generation with XGBoost: case study in Norway. Nordic Machine Intelligence, 6(2), 42-47. https://doi.org/10.5617/nmi.12933
Vancouver
Parviero R, He X, Hashemi N, Topka W, Martinelli G. Prediction of wind power generation with XGBoost: case study in Norway. Nordic Machine Intelligence. 2026;6(2):42-47. doi:10.5617/nmi.12933
BibTeX
@article{Parviero2026,
author = {Parviero, Riccardo and He, Xiulan and Hashemi, Naser and Topka, Wioleta and Martinelli, Gabriele},
title = {Prediction of wind power generation with XGBoost: case study in Norway},
journal = {Nordic Machine Intelligence},
year = {2026},
volume = {6},
number = {2},
pages = {42--47},
doi = {10.5617/nmi.12933}
}
EUVIMED