Multi-Label User Portrait Model for Charging Station Demand: Design and Verification

This study aims to mine user portrait based on data from electric vehicle users' charging behavior, which includes fields such as user ID, starting remaining battery level, ending remaining battery level, charging amount, charging duration, charging start time, charging end time, and charging station ID. Given that the current automotive user portrait outputs system is not well-established, with limited feature dimensions for analysis and portrait outputs that do not adequately support management's comprehensive understanding of user behavior and needs, this research improves upon the traditional RFM model. It constructs a more suitable RFM model for describing electric vehicle charging behavior and introduces a time-decayed label matching method for automotive portrait outputs.