Harnessing Machine Learning Algorithms for Accurate Prediction of CO₂ Emissions in Passenger Vehicles: A Data-Driven Analytical Framework

The escalating global concern over vehicular carbon emissions necessitates advanced predictive models to mitigate environmental degradation. This study develops a machine learning (ML) framework to accurately forecast CO₂ emissions from passenger vehicles using engine specifications, fuel consumption, and vehicle attributes. Employing supervised learning algorithms Linear Regression, Random Forest, Gradient Boosting, and Support Vector Regression the research evaluates their predictive efficacy on a dataset of 1,200 vehicles. Results indicate Gradient Boosting as the most accurate (R² = 0.94, RMSE = 8.34 g/km), with fuel consumption and engine size being the strongest emission determinants. The findings underscore ML’s superiority over traditional emission models, offering actionable insights for policymakers and manufacturers to optimize eco-friendly vehicle designs. The study bridges the gap between data-driven environmental analytics and sustainable transport planning, advocating for standardized emission datasets and interpretable AI tools to enhance regulatory compliance and public awareness. Keywords: CO₂ emissions, machine learning, predictive modeling, passenger vehicles, environmental sustainability, Gradient Boosting.