Prediction of Medium-duration Subway Passenger Flow Volume based on the ARIMA Model

The subway industry has brought about a significant change in transportation by effectively reducing ground traffic congestion, resulting in an increasing demand for predicting subway passenger flow based on historical data. Researchers are constantly striving to improve the diversity and accuracy of data prediction models. This paper examines the daily passenger flow data of Nanjing Metro in 2023 and makes medium-term predictions using the ARIMA model to explore the feasibility and effectiveness of this approach. ADF, ACF, and PACF tests are conducted on the data to ensure that the parameters input into the model can optimize its accuracy. Then the ARIMA model is utilized to fit the data, resulting in a highly accurate parameter model. The results predict the passenger flow of the Nanjing subway in the next ten days, showing that the model's fitted values closely resemble the true values' distribution. By utilizing the ARIMA model, predictions for the next 10 days are made, yielding relatively accurate results. This paper demonstrates that the ARIMA model can be effectively applied to predict subway passenger flow in the medium term, thereby improving the precision of the forecasts.