Prediction of EV Charging Behavior Using Machine Learning

As a key pillar of smart transportation in city applications, electric vehicles (EVs) are becoming increasingly popular for their contribution reducing greenhouse gas emissions. One the challenges, however, is strain on power grid infrastructure that comes with large-scale EV deployment. The solution to this lies utilization scheduling algorithms manage growing public charging demand. Using data-driven tools and machine learning learn behavior can improve algorithms. Researchers have focused using historical data predictions such as departure time energy needs. However, variables weather, traffic, nearby events, which been neglected large extent, perhaps add meaningful representations, provide better predictions. Therefore, paper we propose usage conjunction events predict session duration consumption including random forest, SVM, XGBoost deep neural networks. best predictive performance achieved by an ensemble model, SMAPE scores 9.9% 11.6% consumptions, respectively, improves upon existing works literature. In both predictions, demonstrate significant improvement compared previous work same dataset highlight importance traffic weather information