Abstract

The transportation sector is a significant contributor to global greenhouse gas emissions, with carbon dioxide (CO₂) from vehicles being a primary driver of climate change. Accurate prediction of vehicle CO₂ emissions based on engine and fuel economy characteristics is essential for regulatory compliance, environmental policy, and automotive design optimization. In this research, the EPA Model Year 2026 Fuel Economy Guide dataset, comprising 652 vehicle records with 15 attributes including engine displacement, cylinder count, fuel economy ratings, and CO₂ emission measurements, is utilized to build and evaluate machine learning regression models. Three supervised learning algorithms are implemented: Random Forest Regression, K-Nearest Neighbors (KNN) Regression, and Support Vector Regression (SVR). Each model is trained on 80% of the data and tested on the remaining 20%, with performance evaluated using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), R-squared (R²), and Adjusted R-squared metrics. Additionally, 5-fold cross-validation is employed to assess model robustness across different data partitions. The results demonstrate that SVR with an RBF kernel achieves the highest predictive accuracy with an R² of 0.9988 and MAE of 2.16 grams per mile, followed closely by Random Forest (R² = 0.9978). This study provides a framework for applying machine learning techniques to vehicle emission prediction and highlights the potential of data-driven approaches for supporting environmental sustainability initiatives.

References

• Intergovernmental Panel on Climate Change (IPCC), “Climate Change 2023: Synthesis Report,” Contribution of Working Groups I, II and III to the Sixth Assessment Report, 2023.
• G. Çınarer, M. K. Yeşilyurt, Ü. Ağbulut, Z. Yılbaşı, and K. Kılıç, “Application of various machine learning algorithms in view of predicting the CO₂ emissions in the transportation sector,” Science and Technology for Energy Transition, vol. 79, p. 15, 2024.
• F. Gurcan, “Forecasting CO₂ emissions of fuel vehicles for an ecological world using ensemble learning, machine learning, and deep learning models,” PeerJ Computer Science, vol. 10, e2234, 2024.
• M. H. Memon et al., “Deep learning model based prediction of vehicle CO₂ emissions with eXplainable AI integration for sustainable environment,” Scientific Reports, vol. 15, no. 3602, 2025.
• A. Mohammed, R. Sowah, and E. Annan, “Application of Machine Learning to Predict CO₂ Emissions in Light-Duty Vehicles,” Sensors, vol. 24, no. 24, p. 8219, 2024.
• Ü. Ağbulut, “Forecasting of transportation-related energy demand and CO₂ emissions in Turkey with different machine learning algorithms,” Sustainable Production and Consumption, vol. 29, pp. 141–157, 2022.
• L. Breiman, “Random Forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, 2001.
• T. Cover and P. Hart, “Nearest neighbor pattern classification,” IEEE Transactions on Information Theory, vol. 13, no. 1, pp. 21–27, 1967.
• A. J. Smola and B. Schölkopf, “A tutorial on support vector regression,” Statistics and Computing, vol. 14, no. 3, pp. 199–222, 2004.
• F. Pedregosa et al., “Scikit-learn: Machine Learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011