Convolution Short-Time Passenger Flow Prediction Model of Urban Rail Transit Considering Interpretability

This study focuses on the passenger flow prediction in subway stations, especially in the micro-network scenario. We use Granger causality analysis method and deep learning model to explore the relationship of passenger flow, which improves the accuracy and interpretability of short-term prediction of urban rail transit passenger flow. Based on the analysis of temporal and spatial causality of monitoring points, a convolution short-time passenger flow prediction model (ICCM-URTPFP model) of urban rail transit causality considering interpretability was constructed, and compared with the traditional model, which verified its advantages in passenger flow prediction.