摘要
Abstract
Based on highway ETC toll records,freight vehicle trajectory data,and corporate business registration infor-mation,we have constructed a multi-source feature system covering dimensions such as basic corporate attri-butes,spatio-temporal travel characteristics,and toll contribution.Following data cleaning,standardization,and feature selection,we applied the K-prototype clustering algorithm—suitable for mixed-type data—to cluster and profile freight companies,identifying five typical user groups.By further integrating the travel characteristics and value contributions of different groups,we propose a tiered,differentiated tolling strategy and precision marketing approach,and validate their feasibility through scenario-based simulations.The study demonstrates that refined operations based on user profiling can effectively facilitate the transition of ex-pressways'charging model from"uniform tolling"to"segmented operations",providing decision support for traffic diversion,revenue generation,and the optimal allocation of road network resources.关键词
高速公路/货运企业/用户画像/聚类分析/差异化运营Key words
expressway/freight enterprises/user profiles/cluster analysis/differentiated operation分类
交通工程
