A novel predictive framework for green transportation and EV policy towards sustainable mobility and emission reduction

Abstract

The transportation industry has emerged as the cornerstone of economic development while simultaneously being a major source of greenhouse gas emissions. In response to these increasing environmental concerns, transitioning to green transportation, particularly through the adoption of electric vehicles (EVs), is imperative. Consequently, we propose a supervised machine learning (ML) approach using the Random Forest Regressor algorithm to forecast EV adoption trends, evaluate fuel cost savings, estimate carbon emission reductions, and optimize logistics for sustainability. The resulting model demonstrates excellent regression metrics for MSE and MAE with an R${}^2$ score of 0.998. The study also incorporated real growth rate scenarios based on empirical data and international benchmarks to quantify the impact of financial incentives, the development of charging infrastructure, and public awareness on project EV adoption rates, fuel cost savings, and CO${}_2$ reduction for the years 2024–2034. Although ML-based simulations indicated a stronger standalone impact of charging infrastructure over financial incentives, it was the combined policy scenarios that yielded the most significant outcomes. These results are significant because they establish a scalable model capable of accurately projecting the dynamics of the EV market and environmental outcomes, equipping policymakers and manufacturers with actionable insights to make strategic decisions and contribute meaningfully to the mitigation of carbon emissions.