Official implementation of MR-UFP (Neural Networks 2025): Enhancing urban flow prediction via mutual reinforcement with multi-scale regional information. The paper calls the method MR-UFP; this repository is named MR-UPF.
MR-UFP predicts the inflow and outflow of city grid regions, even with limited training data, by pre-training encoders with spatial-temporal masking and contrastive learning and training flow prediction jointly with a multi-scale region-classification task. On full TaxiBJ and BikeNYC it beats all 13 baselines (RMSE 14.32 and 4.45).
📄 Paper: https://doi.org/10.1016/j.neunet.2024.106900 · 🌐 Paper page with quoted results, FAQ and BibTeX: https://codezx6.github.io/papers/mr-ufp.html
A deep learning framework for spatio-temporal prediction incorporating multi-scale semantic representations and heterogeneous feature fusion mechanisms.
The framework consists of hierarchical encoding modules with cross-modal attention mechanisms and adaptive feature aggregation strategies.
If you use this code, please cite the corresponding paper.
@article{zhang2025mrufp,
title = {Enhancing urban flow prediction via mutual reinforcement with multi-scale regional information},
author = {Zhang, Xu and Cao, Mengxin and Gong, Yongshun and Wu, Xiaoming and Dong, Xiangjun and Guo, Ying and Zhao, Long and Zhang, Chengqi},
journal = {Neural Networks},
year = {2025},
volume = {182},
pages = {106900},
doi = {10.1016/j.neunet.2024.106900},
issn = {0893-6080},
url = {https://doi.org/10.1016/j.neunet.2024.106900}
}