Research on Railway Connection Modes Behavior and Users Profiles Based on Passenger Clustering

Due to the higher requirements of the intelligent and personalized service level in the entire travel process, Maas has become a key research area for scholars. As an important part of the full trip chain, the connection stage plays an significant role in the satisfaction of railway passengers. To provide a service that better cater to individual passengers, this paper proposes an intelligent approach for identifying profiles of railway passengers’ connection preference. First, passengers were clustered into three segments by improved adaptive DBSCAN algorithm based on their personal attributes which captured in the questionnaire. Then, multivariate Logistic regression model is used to fit the parameter values of the passenger group’s selection behavior of connection mode. According to the parameter, user profiles of the three segments are identified and verified: (i) economy-preferred passengers; (ii) convenience-preferred passengers; (iii) time-preferred passengers. This method of profile portraying is able to formulate personalized and differentiated marketing strategies by matching more accurate and efficient connection travel plans for specific railway passengers groups.