AI - Driven Prediction of EV Charging Station Load Demand in Indonesia: A Machine Learning Approach

This study applies ML (Machine Learning) techniques to enhance the efficiency of Electric Vehicle (EV) Charging Station (EVCS) load demand management in Indonesia, leveraging data from the PLN-Mobile application. It addresses the limitations of past research models by incorporating specific user behavior characteristics and EVCS location features into the forecasting model, aiming to provide a more accurate prediction of EVCS load demand. Through a comprehensive Exploratory Data Analysis (EDA) process and the evaluation of six regression models (Decision Tree, Random Forest, Gradient-Boost, XGBoost, Light GBM, and CatBoost) the study concludes that specific user behavior characteristics and EVCS location features had influenced load demand predictions, contributing to the broader discourse on the application of AI and ML in the energy sector. The best accuracy was given at $R^{2}$ 0.958 with CatBoost using 23 features.