Arrival Passenger Volume Prediction Method Based on Bi-LSTM Model at Metropolitan External Transportation Hubs

The metropolitan external transportation hub is a key transport node where urban transportation and inter-city travel intersect. Because there is an uneven distribution of time and space in the arrival passenger flow, we choose the Bi-LSTM model to predict the arrival passenger volume at a metropolitan external transportation hub and use the whale algorithm (WOA) to determine the optimal combination of parameters for the model. This paper used arrival passenger volume data of Beijing South Railway Station from April to May in 2021. The first 70% data was used for training the Bi-LSTM model, and the second 30% was used for prediction. We show that the WOA-Bi-LSTM model had higher prediction accuracy compared with other neural network models. The Bi- LSTM model can accurately predict the trend of arrival passenger volume at external transportation hubs, helping promote the capacity of connecting transportation and service guarantee at the hub.