Mon, May 19, 2025
[Archive]
Hakkari University
Abstract: (109 Views)
The accurate energy consumption prediction for OPEC (Organization of the Petroleum Exporting Countries) member states is vital for strategic planning and policy-making in the global energy market. This study leverages advanced machine learning techniques to forecast energy consumption, utilizing historical data from the U.S. Energy Information Administration (EIA). We applied a variety of machine learning models, including Simple Linear Regression, Gaussian Processes, Multilayer Perceptron (MLP), SMOreg, IBK, Kstar, LWL, Random Subspace, Random Committee, and Random Forest, to the task of predicting energy consumption. The performance of these models was evaluated based on metrics such as R-squared (R²), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Relative Absolute Error (RAE), and Root Relative Squared Error (RRSE). Our results demonstrated that the Random Committee model achieved the highest accuracy in both training (R² = 0.9999, MAE = 0.7411, RMSE = 1.0509, RAE = 1.2199%, RRSE = 1.2399%) and testing phases (R² = 0.9525, MAE = 11.4795, RMSE = 30.6585, RAE = 17.9586%, RRSE = 31.6700%), highlighting its robustness and predictive power. In contrast, the LWL model showed the poorest performance, with significant errors in both phases. The study also highlights the strengths and limitations of each model, with a focus on the applicability of these findings for policymakers and energy analysts. The insights gained from this research underscore the potential of machine learning to enhance energy consumption forecasting, providing a foundation for future studies to build upon. Directions for future research include incorporating additional socio-economic and environmental variables, real-time data, and more advanced machine learning techniques to improve prediction accuracy and reliability further.
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
