Prediction of Passenger Flow in the Associated Region outside Subway Station

This paper aims to accurately predict the passenger flow in the associated region outside subway station. For this purpose, the concept of passenger pool was introduced to abstract the associated region outside the subway station. Considering the features of passenger flow, a nonstationary time series model was established to determine the input and output of the pool. Then, the cumulative residual error was controlled within the confidence interval by the ARIMA model and the deep residual network, aiming to enhance the accuracy and reduce the cumulative error of the model during the calculation of the passenger flow in the associated region. Finally, the proposed model was verified through a case study on Bajiao Amusement Park Station of Beijing Subway Line 1. The results demonstrate the accuracy of the model in predicting the passenger flow aggregation in the associated region outside the station. The research findings provide a decisive basis for warning against abnormal passenger flows, making it possible to realize efficient control of passenger flow.