Advancing Carbon Emission Prediction through Machine Learning: The Impact of Fossil, Nuclear, and Renewable Energies

Accurate prediction of CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions is significant in the development of appropriate strategies to mitigate climate change. This study applies machine learning methodologies for the forecast of CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions using a global dataset from I960 to 2022. The dataset used from Kaggle comprises all the relevant features: power output from fossil fuels, nuclear, and renewables, in addition to GDP and land area. Three machine learning models were constructed and evaluated for their effectiveness in prediction: Random Forest, Gradient Boosting, and Linear Regression. The Random Forest model had the highest accuracy, as demonstrated by its R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.995, which showcases its strength in handling intricate, nonlinear interactions and identifying key variables, such as dependence on fossil fuels. The R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score for Gradient Boosting was 0.987, which shows that this model is similarly good at identifying complex relationships between features. Conversely, Linear Regression underperformed despite its processing efficiency due to its constraints with nonlinear data. This study highlights the efficacy of ensemble models in predicting CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions. It offers practical recommendations for policymakers to formulate strategies to decrease fossil fuel use and promote renewable energy utilization. Future endeavors will include real-time data and investigate hybrid models to enhance temporal forecasting and practical application.