Carbon Prognosticator: A Triple Ensemble Regressor and SHAP Analysis for Enhanced CO₂ Emission Modeling

Tracking carbon emissions from both vehicles and humans is crucial for understanding and mitigating climate change. Accurate prediction models are essential to address the diverse nature of carbon emission datasets. An ensemble regression model combining Random Forest, CatBoost, and Deep Neural Networks has been proposed to effectively handle various types of datasets. Evaluations showed that the model achieved an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex> score of 99.717% on the CO2 Emission (Vehicles) dataset with a 70–30 train-test split and an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex> score of 98.425% on the Carbon Footprint (Humans) dataset with an 80–20 split. These results underscore the model's capability to provide robust and accurate predictions across different carbon emission contexts, enhancing environmental analytics.