<![CDATA[Electric vehicle (EV) adoption is increasing with increasing awareness of clean energy and environmental concerns. However, range anxiety, the uncertainty in estimating driving range, remains a major barrier. This study aims to develop a predictive model for EV range based on technical specifications to provide more accurate estimates and alleviate user concerns. A Machine Learning approach is applied, using Random Forest for feature selection and XGBoost as the primary prediction algorithm. The dataset consists of 478 EV records with 22 attributes, including battery capacity, efficiency, dimensions, and speed. Key features affecting range prediction include battery_capacity_kWh, efficiency_wh_per_km, and height_mm. The XGBoost model demonstrates strong predictive performance with an R² of 0.978, an MAE of 10.555, and an RMSE of 15.180. These results suggest that combining Random Forest and XGBoost offers a promising solution to improve the accuracy of EV range estimation, potentially reducing range anxiety and supporting wider EV adoption]]>
Copyrights © 2026
