User Matching Algorithm Based on Repeated Transfer Behavior

Trip data contains important information about travelers' behaviors and patterns, but it has not been fully explored and utilized. One of the reasons that cause this problem is that these data need to be anonymized when shared and analyzed, which makes it difficult to link trips together to form trip tour. As a result, this study intends to propose a method to match users across transportation modes (bikeshare and metro in this study) making full use of the characteristics of the transfer behavior, and the matching results are further pruned based on time and space constraints. The results show that there is a strong correlation between bikeshare travel and metro travel, which is consistent with the results of previous studies. Two empirical formulas are obtained by fitting bikeshare users with different trip times to the matching results. The goodness of fit is 0.981 and 0.865, respectively, and the matching effect is good. At the same time, it also shows that the algorithm in this paper can better handle the matching of users with transfer and commuting behavior. This algorithm can provide important support for transportation planning and operation management of the urban transportation system.