Synthetic Data-Driven Analysis of Electric Vehicle Manufacturing in India Using Machine Learning

Abstract

The global trend to Electric Vehicles is crucial in the minimization of carbon emissions in the transportation field. Nonetheless, countries like India have serious analytical difficulties in the change due to the paucity and irregularity of long-term manufacturing data. The current work attempts to fill this gap by proposing a Machine Learning model based on the synthetic data to forecast the future trends in the production of EVs in the country. An artificial dataset over the years 2010-2025 was thoroughly constructed based on fifteen selected variables that reflect governmental policy, supply-chain capacity, and significant economic indicators. A high-dimensional, detailed parameter space feature base was trained on these parameters to create a system based on a Multi-Output Random Forest Regressor that can simultaneously predict two parameters: EV production performance and corresponding transport-sector CO₂ emissions. Projections of the years 2026-2035 have been created with the use of this setup under the three policy and market conditions, namely, Conservative, Base, and Accelerated. The analysis of these scenarios shows that the Accelerated case provides the best results, predicting the sharp increase of the EV production and a total decrease of CO₂ of approximately 58 Mt over the ten years. On the whole, the created model plays the role of an effective decision-support tool for both policymakers and industrial planners.