Modeling CO2 Emissions in Morocco: A Support Vector Regression Approach

This research employs Support Vector Regression (SVR) to analyze CO2 emissions in Morocco, incorporating a dataset spanning 1965 to 2021 and featuring energy consumption, GDP, industrial production, population and urbanization—categorized as economical, demographical, and energetical factors. The SVR model, optimized via Grid Search, predicts emissions accurately, yielding an R-squared of 0.99, MSE of 0.04, and MAE of 0.19. With 80% of data allocated to training, the model aligns effectively with historical trends, demonstrating proficiency in predicting emissions beyond 2010. The study's implications for policy, industry, and sustainability underscore the importance of renewable energy and sustainable practices in Morocco's environmental goals. These findings not only highlight the critical role of ML models in CO2 prediction but also offer actionable insights for informed decision-making and environmental management.