Short-term metro passenger flow forecasting via decoupled spatio-temporal modeling with a diffusion-convergence graph and Swin-Enhanced LSTM
- 1Department of Computer and Software Engineering, Huaiyin Institute of Technology, Huaian, Jiangsu, China
- 2Trusted Firmware and Intelligent Software Laboratory, Huaian, Jiangsu, China
- 3School of Transportation, Southeast University, Nanjing, Jiangsu, China
Abstract
Short‑term metro passenger flow forecasting—predicting station‑level inflow and outflow volumes is challenging, mainly due to the complexity of multimodal spatial couplings and the difficulty of modeling long‑range temporal dependencies while retaining local variations. Therefore, this paper proposes an Integrated Spatio-Temporal LSTM (IST-LSTM). IST-LSTM adopts a dual-branch architecture in which a diffusion branch explicitly models OD-driven propagation and a convergence branch captures structural aggregation among stations; these branches are coupled through a Diffusion and Convergence Graph (DCG) that fuses structural links, inter-station distances, and behavior-driven OD diffusion into a unified spatial representation. To strengthen temporal representation, we refine the LSTM gating mechanism by integrating a Swin Transformer module, enabling the model to better represent long-range dependencies while retaining sensitivity to local, fine-grained temporal patterns. To evaluate model generalization under dense and heterogeneous urban conditions, we construct the NJMetro dataset from Nanjing AFC smart-card transaction data, which reflects inner-ring high-density travel dynamics and pronounced spatio-temporal heterogeneity; this dataset is used alongside the public HZMetro dataset. IST-LSTM reduces RMSE by 9.19% on NJMetro relative to GCN-SBULSTM and achieves an average MAE improvement of 9.62% over Graph WaveNet on HZMetro and 8.68% versus LSTM on NJMetro. On HZMetro, it also achieves competitive RMSE performance relative to PB-GRU, with notable advantages at 15- and 30-minute horizons. Overall, the results indicate that the proposed design more effectively captures the coupled spatial and temporal mechanisms underlying metro passenger flow.
Related
- Diffusion-enhanced spatio-temporal models for estimating spatial expansion
- Diffusion tensor imaging of whole human brains during long-term formaldehyde fixation: Temporal evolution of diffusion parameters, post-mortem conditions, and dependence on tissue structure
- A multilayer drift‐diffusion broadcast model: Converting temporal evidence into response timing
- Constraints on the G1/S transition pathway may favor selection of multicellularity as a passenger phenotype
- Species-specific drivers of genetic diversity are decoupled from landscape connectivity and plant community diversity
BibTeX
@article{feng2026shortterm,
title = {{Short-term metro passenger flow forecasting via decoupled spatio-temporal modeling with a diffusion-convergence graph and Swin-Enhanced LSTM}},
author = {Yixin Feng and Shangbing Gao and Hao Wang and Anming Bai and Gang Ren and Yuanyuan Wang and Guxue Gao and Jiahuan Ren},
journal = {PLOS One},
year = {2026},
month = oct,
doi = {10.1371/journal.pone.0359409},
url = {https://doi.org/10.1371/journal.pone.0359409},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Short-term metro passenger flow forecasting via decoupled spatio-temporal modeling with a diffusion-convergence graph and Swin-Enhanced LSTM
AU - Yixin Feng
AU - Shangbing Gao
AU - Hao Wang
AU - Anming Bai
AU - Gang Ren
AU - Yuanyuan Wang
AU - Guxue Gao
AU - Jiahuan Ren
PY - 2026
DA - 2026/10/06
JO - PLOS One
DO - 10.1371/journal.pone.0359409
UR - https://doi.org/10.1371/journal.pone.0359409
AB - Short‑term metro passenger flow forecasting—predicting station‑level inflow and outflow volumes is challenging, mainly due to the complexity of multimodal spatial couplings and the difficulty of modeling long‑range temporal dependencies while retaining local variations. Therefore, this paper proposes an Integrated Spatio-Temporal LSTM (IST-LSTM). IST-LSTM adopts a dual-branch architecture in which a diffusion branch explicitly models OD-driven propagation and a convergence branch captures structural aggregation among stations; these branches are coupled through a Diffusion and Convergence Graph (DCG) that fuses structural links, inter-station distances, and behavior-driven OD diffusion into a unified spatial representation. To strengthen temporal representation, we refine the LSTM gating mechanism by integrating a Swin Transformer module, enabling the model to better represent long-range dependencies while retaining sensitivity to local, fine-grained temporal patterns. To evaluate model generalization under dense and heterogeneous urban conditions, we construct the NJMetro dataset from Nanjing AFC smart-card transaction data, which reflects inner-ring high-density travel dynamics and pronounced spatio-temporal heterogeneity; this dataset is used alongside the public HZMetro dataset. IST-LSTM reduces RMSE by 9.19% on NJMetro relative to GCN-SBULSTM and achieves an average MAE improvement of 9.62% over Graph WaveNet on HZMetro and 8.68% versus LSTM on NJMetro. On HZMetro, it also achieves competitive RMSE performance relative to PB-GRU, with notable advantages at 15- and 30-minute horizons. Overall, the results indicate that the proposed design more effectively captures the coupled spatial and temporal mechanisms underlying metro passenger flow.
SN - 1932-6203
ER -