Akademik Veri Yönetim Sistemi
Balkan and Near Eastern Journal of Social Sciences, cilt.0, sa.11, ss.168-171, 2025 (Hakemli Dergi)
Road transportation is a major source of atmospheric carbon dioxide (CO2), posing a significant threat to global climate stability and environmental sustainability. According to the International Energy Agency, the transport sector is responsible for approximately one-quarter of worldwide energy-derived CO2 emissions, underscoring its role in the global carbon footprint. The continuous rise in the number of passenger vehicles has intensified concerns regarding environmental degradation, deteriorating air quality, and associated public health risks. This study presents a machine learning based framework for predicting vehicle CO2 emissions using a publicly available dataset. Three regression models were implemented to provide a comparative performance evaluation. To improve model interpretability, feature importance analysis was applied using the Shapley Additive explanations method. The experimental findings indicate that ensemble-based models outperform the linear regression approach, achieving superior predictive accuracy across all evaluation metrics. Overall, the proposed framework improves reliability in CO2 emission prediction, while offering actionable insights for policymakers, manufacturers, and consumers aiming to support low-carbon transportation.