Research on the Metro Ridership Forecasting based on ARIMA Model

Abstract

With the increasing coverage of subways in cities, people travel more than just by bus or walking. Nowadays, subway stations have become sites with high population density in the city. Due to the increasing travel demand of residents. Excessive congestion in subway stations will lead to inconvenience and even accidents. The subway passenger flow prediction can avoid many potential problems. Passenger flow prediction can not only optimize the subway scheduling, but also help the operating company to fully prepare for the security work during the rush hours. Data from February 1,2023 to December 31,2023 are from Xi'an Transportation Platform. This research used the ARIMA model for mid-term passenger flow prediction, and the data from February to October were used as the training set. The first order of the data is differential to ensure the stability of the data. The values of p and q in the model were determined using the autocorrelation function, and the most appropriate combinations of p and q were selected by BIC. The ARIMA model was built using (p, d, q), and predicted passenger traffic in November and December. Finally, the feasibility of predicting the medium-term subway passenger flow is determined by comparing the predicted value with the true value.