Forecast of Short-Term Passenger Flow of Urban Railway Stations Based on Seasonal ARIMA Model

Abstract

Forecast on short-term passenger flow of urban rail transit is the key to network operation and management, and the basis of passenger flow organization and train optimal allocation. In this chapter, a predictive model of passenger flow entering and departing is constructed, based on the model of seasonal Autoregressive Integrated Moving Average (ARIMA). First, outliers of passenger flow in time series were replaced by the linear interpolation method; Second, two methods, after considering weather conditions and air quality, are used for passenger flow forecast respectively. One is seasonal differencing, the other is by adding working day attributes as dummy variables. Third, the method of least squares was used to estimate the weight, thus a combined forecasting model for time series was constructed. After, the model has been calibrated and validated by the historical passenger flow data collected by AFC system of Beijing Metro: the error is less than 5%. This model not only considered dummy variables such as weather conditions, air quality, and working day attributes, but also quantified their impact for passenger flow. The results show that the prediction model has high accuracy.