Passenger Flow Prediction of Tianjin Metro Line 3 under Time Series Clustering

Abstract

Taking Tianjin Metro Line 3 as an example, this study mainly explored the influence of forecasting methods on the accuracy of short-term passenger flow forecasting for rail transit. Based on the station's inbound and outbound passenger flow attributes, the time sequence clustering method was applied to classify urban rail transit stations into four categories, and extreme gradient boosting (XG Boost), back propagation (BP) and autoregressive moving average (ARMA) models were used to predict the short-term passenger flow of each type of stations, and the results were compared. The results show that the methods with the highest accuracy in predicting the passenger flow of the four types of stations are XG Boost, ARMA, ARMA and BP, and XG Boost shows a greater advantage in prediction time.