Fuel Efficiency and Emission Prediction Using Auto MPG and EPA Data: A Machine Learning and Deep Learning Approach

Year 2025, Volume: 6 Issue: 2, 30 - 38

Zeynep Çeken , Fahriye Gemci

https://doi.org/10.55195/jscai.1774127

Abstract

This study consists of two stages. In the first stage, the Auto MPG dataset, and in the second stage, the Fuel Economy dataset published by the United States Environmental Protection Agency (EPA) are utilized to classify vehicle fuel efficiency and predict carbon emissions. In this context, multi-layered and multi-functional artificial intelligence and machine learning-based models capable of solving both classification and regression problems were applied. Comparative analyses were conducted by testing different deep learning architectures such as Random Forest Classifier, Support Vector Classifier (SVC), Logistic Regression, Artificial Neural Network (ANN), and Convolutional Neural Network (CNN).
In the first stage of this project, it was aimed to model the MPG (Miles Per Gallon) value, which represents fuel consumption efficiency, using both regression and classification methods based on vehicles’ technical parameters. In the second stage, the classification model focuses on predicting vehicle categories such as A+, A, B, C, and D based on the “Fuel Economy Score,” while the regression model aims to accurately predict the “Tailpipe CO2 in Grams/Mile (FT1)” value. The performance of the models was evaluated using various metrics such as accuracy, ROC curves, and mean error [15].
It was observed that the designed XGBoost + CNN model achieved highly successful results. This indicates that CNN models can be effectively applied not only in image processing tasks but also in diverse datasets.
The results reveal that both deep learning and traditional models provide high accuracy in vehicle efficiency prediction and environmental impact analysis. In this respect, the study offers significant contributions to intelligent transportation systems and sustainable environmental policies.

Keywords

Fuel efficiency of vehicles , Carbon emissions of vehicles , Convolutional Neural Network , Random Forest , Xgboost

Thanks

We thank Kahramanmaraş Sütçü İmam University

References

• United States Environmental Protection Agency (EPA), “Fuel economy data (1984–present),” [Online]. Available: https://www.fueleconomy.gov/feg/. Accessed June 16, 2025.
• UCI Machine Learning Repository, “Auto MPG dataset,” [Online]. Available: https://archive.ics.uci.edu/ml/datasets/auto+mpg. Accessed June 16, 2025.
• 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.
• C. Cortes and V. Vapnik, “Support-vector networks,” Machine Learning, vol. 20, no. 3, pp. 273–297, 1995.
• T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining (KDD), pp. 785–794, 2016.
• D. W. Hosmer and S. Lemeshow, Applied Logistic Regression. New York: Wiley, 2000.
• F. Rosenblatt, “The perceptron: A probabilistic model for information storage and organization in the brain,” Psychological Review, vol. 65, no. 6, pp. 386–408, 1958.
• Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no. 11, pp. 2278–2324, 1998.
• I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA: MIT Press, 2016.
• A. M. Foley and A. G. Olabi, “Renewable energy technology and the environment: a review,” Renewable and Sustainable Energy Reviews, vol. 79, pp. 1321–1340, 2017.
• X. Wu, V. Kumar, J. R. Quinlan, et al., “Top 10 algorithms in data mining,” Knowledge and Information Systems, vol. 14, no. 1, pp. 1–37, 2008.
• T. Fawcett, “An introduction to ROC analysis,” Pattern Recognition Letters, vol. 27, no. 8, pp. 861–874, 2006.
• J. A. Hanley and B. J. McNeil, “The meaning and use of the area under a receiver operating characteristic (ROC) curve,” Radiology, vol. 143, no. 1, pp. 29–36, 1982.
• K. H. Zou, A. J. O’Malley, and L. Mauri, “Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models,” Circulation, vol. 115, no. 5, pp. 654–657, 2007.
• S Akhatkulov,I Yalgoshev, Z. and Urinboyev, ”VehicleCO2 Emission Prediction Using Deep Learning and Ensemble Machine Learning Methods,” In 2025 International Russian Automation Conference (RusAutoCon) (pp. 819-824). IEEE.
• T Ji, K Li, Q Sun, and Z Duan,”Urban transport emission prediction analysis through machine learning and deep learning techniques” Transportation Research Part D: Transport and Environment, vol. 135, no. 104389, pp. 654–657,2024