The 22nd COTA International Conference of Transportation Professionals (CICTP 2022)

Short-term passenger flow prediction plays an important role in the guidance, control, and management of intelligent transportation systems. Aiming at the problems of unclear influencing factors and low prediction accuracy of the current short-term prediction methods of passenger flow, this paper proposes a passenger flow prediction method based on gradient boosting. Based on the spatio-temporal correlation passenger flow, the features that may affect passenger flow are extracted from the inbound AFC (Automatic Fare Collection System) data of urban rail transit. The data set is aggregated by the passenger flow every ten minutes. Finally, the LightGBM (light gradient boosting machine) model is established to realize efficient and accurate short-term passenger flow prediction. Experimental verification based on the AFC inbound data set of Nanjing rail transit shows that the prediction accuracy of the LightGBM model is higher than that of the Fbprophet model.