Identification of City Couriers Based on Mobile Phone Data

An identification method of urban express based on the naive Bayesian classifier (NBC) was proposed in this paper. Firstly, the rules of express delivery personnel phone signaling was researched. And then, based on the mobile phone communication signaling data in Beijing, the Bayesian probabilistic relations were established between the interviewer’s category variable (couriers/non-couriers) and the bus travel information which were contained in both questionnaire and mobile phone signaling data. On the basis of this, the Bayesian model was constructed and its training was carried out. Finally, the accuracy of the calibrated model was tested by using the sample data which was not involved in training, and the test showed that the success rate of the courier prediction reached 88.3%. It is shown that the method has high accuracy, which can meet the demand of practical application.