Electric Vehicle Adoption and Type Prediction in India Using Machine Learning

Abstract: As the world switches to sustainable energy, there has been an increase in the adoption of Electric Vehicles (EVs) in India due to the combined efforts of government policy and growing consumer awareness of environmentally friendly products. However, stakeholders in the industry face challenges in predicting future demand for EVs and recognizing the leading types of vehicles in different regions. In this study, a dual-model approach is proposed in order to forecast future demand and predict the leading vehicle types: a Prophet-type model that provides forecasted monthly EV sales for each state, and a Random Forest classifier that predicts the leading vehicle type based on historical data. The dual models are also built into a complete stack Flask-Android prototype that provides an average MAPE for forecasted sales of 6.56% and an accuracy of 55.3% for predicted vehicle types.

Keywords: Electric Vehicles, Machine Learning, Prophet, Random Forest, Sales Forecasting, Time-Series Analysis, Flask, Android Application.