This study focuses on utilizing multi-source data for metro passenger flow prediction in a practically and computationally efficient way. To further improve the accuracy and reliability of passenger prediction, a deep learning-based framework is proposed in this paper considering the practical interference of neighbor spatial dependence and train events, namely a multi-disturbance spatial-temporal causal convolution network (MD-STCN). Specifically, the passenger flow features and train event features of adjacent areas are constructed as initial input to the approach through feature modelling and data fusion. Furthermore, a two-stage spatio-temporal feature selection algorithm is developed to obtain a compact input. Finally, four datasets from Shanghai Metro are tested to verify the validity and feasibility of our MD-STCN.
MD-STCN: A deep learning-based architecture considering multivariate disturbances for metro passenger flow prediction
2022-10-08
497600 byte
Conference paper
Electronic Resource
English
A deep learning and ensemble learning based architecture for metro passenger flow forecast
Wiley | 2023
|A deep learning and ensemble learning based architecture for metro passenger flow forecast
DOAJ | 2023
|A Hybrid Spatiotemporal Deep Learning Model for Short-Term Metro Passenger Flow Prediction
DOAJ | 2020
|Metro Passenger Volume Prediction
DataCite | 2024
|