A Comparison Study of Short-Term Passenger Flow Forecast Model of Rail Transit

Short-term passenger flow forecasting is an important data support for urban rail transit operation planning; it improves passenger flow organization, gives early-warning for oversaturation, and improves transportation service and safety. Many forecast models have been proposed to forecast passenger flow dynamics. Different models are based on different theoretical backgrounds with different characteristics; moreover, the dynamic of rail transit passenger flow is different in different stations due to land use and location. Six models, namely MA, ARIMA, SARIMA, BPNN, WNN, and SVM models are used to make short-term predictions of passenger flow within five working days in six Beijing stations. After comparative analysis, prediction accuracies of nonlinear models are relatively higher under a similar model input structure. In addition, this work confirms the existence of strong periodicity and stability of passenger entry flow and suggests that these characteristics should be considered to improve prediction accuracy when constructing models in further research.