MMHFormer: Multi-Source and Multi-View Hierarchical Transformer for Traffic Flow Prediction
Source abstract
Abstract
Traffic flow prediction is a vital component of Intelligent Transportation Systems (ITSs), playing a key role in proactive traffic management and the optimization of urban mobility. However, the complex spatial–temporal dependencies, dynamic variations, and external factors in traffic networks present significant challenges for accurate predictions. In this paper, we propose MMHFormer, a novel multi-source, multi-view hierarchical Transformer model specifically designed for traffic flow prediction. MMHFormer incorporates three key innovations: (1) a multi-source gated embedding layer that integrates diverse multidimensional inputs, including spatial Laplacian embeddings, temporal periodic embeddings, and traffic occupancy embeddings, to better capture the complex dynamics of traffic conditions; (2) a hierarchical multi-view spatial attention module that models global, local, and dynamic similarity-based spatial dependencies, effectively addressing the spatial heterogeneity of traffic flows; (3) a hierarchical two-stage temporal attention mechanism that captures global temporal dependencies while adapting to node-specific temporal variations. Extensive experiments conducted on four benchmark traffic datasets demonstrate that MMHFormer outperforms state-of-the-art methods, achieving significant improvements in prediction accuracy.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Traffic Prediction and Management Techniques · Traffic control and management · Human Mobility and Location-Based Analysis
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Computer Science and Engineering, Chongqing University of Science and Technology, Chongqing 400054, China
School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China
References (39)
Show source-provided references
1. Liu, H., Zhu, C., Zhang, D., and Li, Q. (2023). Multi-Scale Spatial-Temporal Recurrent Networks for Traffic Flow Prediction. arXiv.
2. Kim, K., Jin, S., Ko, S., and Choo, J. (2020, January 19–23). Stgrat: A spatio-temporal graph attention network for traffic forecasting. Proceedings of the International Conference on Information and Knowledge Management, Online.
3. Modeling and Forecasting Vehicular Traffic Flow as a Seasonal ARIMA Process: Theoretical Basis and Empirical Results10.1061/(ASCE)0733-947X(2003)129:6(664) ↗
4. Integrating granger causality and vector auto-regression for traffic prediction of large-scale WLANs
5. Hybrid machine intelligent SVR variants for wind forecasting and ramp events10.1016/j.rser.2019.04.002 ↗
6. The Vanishing Gradient Problem During Learning Recurrent Neural Nets and Problem Solutions10.1142/S0218488598000094 ↗
7. Chung, J., Gulcehre, C., Cho, K., and Bengio, Y. (2014). Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling. arXiv.
8. A fast human action recognition network based on spatio-temporal features10.1016/j.neucom.2020.04.150 ↗
9. Defferrard, M., Bresson, X., and Vandergheynst, P. (2016, January 5–10). Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering. Proceedings of the Advances in Neural Information Processing Systems, Barcelona, Spain.
10. T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction10.1109/TITS.2019.2935152 ↗
11. Hub-hub connections matter: Improving edge dropout to relieve over-smoothing in graph neural networks10.1016/j.knosys.2023.110556 ↗
12. Hybrid deep learning models for traffic prediction in large-scale road networks10.1016/j.inffus.2022.11.019 ↗
13. Short-Term Prediction of Traffic Volume in Urban Arterials10.1061/(ASCE)0733-947X(1995)121:3(249) ↗
14. A novel wavelet-SVM short-time passenger flow prediction in Beijing subway system10.1016/j.neucom.2015.03.085 ↗
15. Sun, B., Cheng, W., Goswami, P., and Bai, G. (2017, January 3–6). Flow-aware WPT k-nearest neighbours regression for short-term traffic prediction. Proceedings of the 2017 IEEE Symposium on Computers and Communications (ISCC), Heraklion, Greece.
16. Liu, Y., Zheng, H., Feng, X., and Chen, Z. (2017, January 11–13). Short-term traffic flow prediction with Conv-LSTM. Proceedings of the 2017 9th International Conference on Wireless Communications and Signal Processing (WCSP), Nanjing, China.10.1109/WCSP.2017.8171119 ↗
17. Lv, Z., Xu, J., Zheng, K., Yin, H., Zhao, P., and Zhou, X. (2018, January 13–19). LC-RNN: A deep learning model for traffic speed prediction. Proceedings of the 27th International Joint Conference on Artificial Intelligence, IJCAI’18, Stockholm, Sweden.10.24963/ijcai.2018/482 ↗
18. Yu, B., Yin, H., and Zhu, Z. (2018, January 13–19). Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecasting. Proceedings of the 27th International Joint Conference on Artificial Intelligence, IJCAI’18, Stockholm, Sweden.10.24963/ijcai.2018/505 ↗
19. Attention Based Spatial-Temporal Graph Convolutional Networks for Traffic Flow Forecasting
20. Wu, Z., Pan, S., Long, G., Jiang, J., and Zhang, C. (2019). Graph WaveNet for Deep Spatial-Temporal Graph Modeling. arXiv.10.24963/ijcai.2019/264 ↗
21. BAI, L., Yao, L., Li, C., Wang, X., and Wang, C. (2020, January 6–12). Adaptive Graph Convolutional Recurrent Network for Traffic Forecasting. Proceedings of the Advances in Neural Information Processing Systems, Virtual.
22. Hierarchical Graph Convolution Network for Traffic Forecasting
23. Dynamic multi-graph neural network for traffic flow prediction incorporating traffic accidents10.1016/j.eswa.2023.121101 ↗
24. Static-dynamic collaborative graph convolutional network with meta-learning for node-level traffic flow prediction10.1016/j.eswa.2023.120333 ↗
25. Multi-view dynamic graph convolution neural network for traffic flow prediction10.1016/j.eswa.2023.119779 ↗
26. Zhong, W., Suo, Q., Jia, X., Zhang, A., and Su, L. (2021, January 7–10). Heterogeneous Spatio-Temporal Graph Convolution Network for Traffic Forecasting with Missing Values. Proceedings of the 2021 IEEE 41st International Conference on Distributed Computing Systems (ICDCS), Washington, DC, USA.10.1109/ICDCS51616.2021.00073 ↗
27. A heterogeneous traffic spatio-temporal graph convolution model for traffic prediction10.1016/j.physa.2024.129746 ↗
28. Wu, M., Lin, Y., Jiang, T., and Weng, W. (2025, January 6–11). MHGNet: Multi-Heterogeneous Graph Neural Network for Traffic Prediction. Proceedings of the ICASSP 2025—2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Hyderabad, India.10.1109/ICASSP49660.2025.10890207 ↗
29. Learning Dynamics and Heterogeneity of Spatial-Temporal Graph Data for Traffic Forecasting10.1109/TKDE.2021.3056502 ↗
30. PDFormer: Propagation Delay-Aware Dynamic Long-Range Transformer for Traffic Flow Prediction
31. LCDFormer: Long-term correlations dual-graph transformer for traffic forecasting10.1016/j.eswa.2024.123721 ↗
32. STGAFormer: Spatial–temporal Gated Attention Transformer based Graph Neural Network for traffic flow forecasting10.1016/j.inffus.2024.102228 ↗
33. DDGformer: Direction- and distance-aware graph transformer for traffic flow prediction10.1016/j.knosys.2024.112381 ↗
34. Li, Y., Yu, R., Shahabi, C., and Liu, Y. (2018). Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting. arXiv.
35. Wu, Z., Pan, S., Long, G., Jiang, J., Chang, X., and Zhang, C. (2020, January 6–10). Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD ’20, Virtual.10.1145/3394486.3403118 ↗
36. GMAN: A Graph Multi-Attention Network for Traffic Prediction
37. Spatial-Temporal Fusion Graph Neural Networks for Traffic Flow Forecasting
38. Shao, Z., Zhang, Z., Wang, F., Wei, W., and Xu, Y. (2022, January 17–21). Spatial-Temporal Identity: A Simple yet Effective Baseline for Multivariate Time Series Forecasting. Proceedings of the 31st ACM International Conference on Information & Knowledge Management, CIKM ’22, Atlanta, GA, USA.10.1145/3511808.3557702 ↗
39. Generic Dynamic Graph Convolutional Network for traffic flow forecasting10.1016/j.inffus.2023.101946 ↗