Short-term Electricity Price Forecasting Using Interpretable Hybrid Machine Learning Models

• Published in: 2023 IEEE IAS Global Conference on Renewable Energy and Hydrogen Technologies (GlobConHT) pp. 1 - 6
• Main Authors: Mubarak, Hamza, Ahmad, Shameem, Hossain, Al Amin, Horan, Ben, Abdellatif, Abdallah, Mekhilef, Saad, Seyedmahmoudian, Mehdi, Stojcevski, Alex, Mokhlis, Hazlie, Kanesan, Jeevan, Becherif, Mohamed
• Format: Conference Proceeding
• Language: English
• Published: IEEE 11.03.2023
• Subjects: CatBoost; electricity price forecasting; hybrid models; Hydrogen; Machine learning; Mean square error methods; Measurement; Predictive models; Renewable energy sources; Support vector machines; time series; XGBoost
• DOI: 10.1109/GlobConHT56829.2023.10087707

In this paper, a combination of single and hybrid Machine learning (ML) models were proposed to forecast the electricity price one day ahead for the Nord Pool spot electricity market. The proposed models were evaluated based on performance metrics, such as Root Mean Square Error (RMSE), Mean Square Error (MSE), and Mean Absolute Error (MAE). Further, a model interpretation by employing SHapley Additive exPlanations (SHAP) framework to show the impact of each feature in the forecasting output. Based on the SHAP, the lag electricity price EP(t-1) impacts the forecast result most, followed by EP(t-2) and time stamp, respectively. Finally, the results show that hybrid models performed better than single ones, where the LR-CatBoost model surpassed other models and attained 7.94 and 10.49, which are the lowest values of MAE and RMSE respectively. Moreover, the kNN and SVM models performed poorly, achieving the highest RMSE values of 12.88 and 12.39, respectively.