EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Hybrid Stacked Ensemble Framework for Electricity Demand Prediction in Urban Setups

Ashwani Kharola, Nitin Kumar, Ruby Pant, Ajay Kumar, Purnendu Shekar Pandey, Jeewan Singh, Harvinder Singh, Abhijit Bhowmik, Yohannes Mengist

International Journal of Energy Research · 2026-01 · 2026 (1)

DOI: 10.1155/er/4160378 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Energy Research, online erschienen am 03.08.2026. Verfasst von Ashwani Kharola und 8 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This article investigates various machine learning, deep learning, and transformer‐based stacked ensemble techniques for electricity demand (ED) prediction in urban environments. The base learners include Extreme Gradient Boosting (XGB), adaptive boosting (ADB), categorical boosting (CATB), long short‐term memory (LSTM), bidirectional LSTM (Bi‐LSTM), and an inverted transformer (I‐TRAN)‐based attention mechanisms. The combination of these base models feeds into a sophisticated random forest (RF) meta‐learner. Feature importance was assessed using SHapley Additive exPlanations (SHAP), while hyperparameters were optimized via the Optuna framework. Utilizing a 20‐fold time series cross‐validation (TSCV) scheme alongside a comprehensive suite of error metrics, the results confirmed that the proposed Bi‐LSTM + RF stacked ensemble framework delivers the best predictive performance. The Bi‐LSTM + RF architecture achieved a perfect rank of 1 across all evaluation metrics, yielding an exceptional root mean square error (RMSE) of 107.504 kW, mean absolute error (MAE) of 58.551 kW, Theil’s U2 of 0.187, a symmetric mean absolute percentage error (SMAPE) of 0.890%, and a dominant explanatory power with an R 2 value of 0.9871, drastically outperforming all baseline and alternative ensemble configurations.

Zitieren

APA

Kharola, A., Kumar, N., Pant, R., Kumar, A., Pandey, P. S., Singh, J., Singh, H., Bhowmik, A., & Mengist, Y. (2026). Hybrid Stacked Ensemble Framework for Electricity Demand Prediction in Urban Setups. International Journal of Energy Research, 2026(1). https://doi.org/10.1155/er/4160378

Vancouver

Kharola A, Kumar N, Pant R, Kumar A, Pandey PS, Singh J, et al. Hybrid Stacked Ensemble Framework for Electricity Demand Prediction in Urban Setups. International Journal of Energy Research. 2026;2026(1). doi:10.1155/er/4160378

BibTeX
@article{Kharola2026,
  author = {Kharola, Ashwani and Kumar, Nitin and Pant, Ruby and Kumar, Ajay and Pandey, Purnendu Shekar and Singh, Jeewan and Singh, Harvinder and Bhowmik, Abhijit and Mengist, Yohannes},
  title = {Hybrid Stacked Ensemble Framework for Electricity Demand Prediction in Urban Setups},
  journal = {International Journal of Energy Research},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/er/4160378}
}
Kostenlos starten →