Predicting energy consumption in Mexico: Integrating environmental, economic, and energy data with machine learning techniques for sustainable development

Accurate energy consumption prediction is crucial for developing effective government energy policies. This study introduces a novel machine learning-based approach for reliable energy consumption prediction to provide stability in forecasting energy usage. This study aims to investigate the predictive capabilities of advanced hybrid machine learning models, specifically CatBoost and SVR hybrids, in accurately forecasting energy consumption based on key variables such as carbon dioxide emissions, global temperature, and oil production. This study investigates the prediction of energy consumption in Mexico, considering its significance for environmental conservation and economic development. Accurate prediction models are developed by using advanced machine learning algorithms such as CatBoost and SVR, along with optimization techniques including Arch-OA, PSO, and ALO. The analysis reveals strong correlations between energy consumption and various factors, emphasizing the importance of considering parameters such as carbon dioxide emissions and GDP per capita. Results demonstrate the superior performance of CatBoost over SVR, with hybrid models incorporating SVR showing notable improvements in accuracy. These findings highlight the efficacy of advanced machine learning and optimization algorithms in accurately forecasting energy demand, thereby facilitating informed policymaking for sustainable development.