Urban road traffic congestion index prediction based on a hybrid LightGBM-LSTM model

Sha Liu; Jian-jie Gao; Yi-lin Hong; Jun-chao Zhou

doi:10.61089/aot2025.zf4t0674

Authors

Sha Liu Sichuan Provincial Key Laboratory of Intelligent Policing, Sichuan Police College, Luzhou, China Author https://orcid.org/0009-0003-8690-8748
Jian-jie Gao Sichuan Provincial Key Laboratory of Intelligent Policing, Sichuan Police College, Luzhou, China Author https://orcid.org/0009-0000-2567-3373
Yi-lin Hong Sichuan Provincial Key Laboratory of Intelligent Policing, Sichuan Police College, Luzhou, China Author https://orcid.org/0009-0002-8036-5204
Jun-chao Zhou School of Mechanical Engineering, Sichuan University of Science ＆ Engineering, Sichuan Zigong, China Author https://orcid.org/0000-0002-5747-3517

DOI:

https://doi.org/10.61089/aot2025.zf4t0674

Keywords:

traffic prediction, hybrid model, LightGBM, LSTM, traffic congestion index

Abstract

Accurate and timely prediction of the urban traffic congestion index (TCI) is crucial for implementing proactive traffic management and alleviating urban congestion. To address the limitations of single models in capturing both complex temporal dependencies and high-dimensional feature interactions, this paper proposes a novel hybrid prediction framework that synergistically integrates a Long Short-Term Memory (LSTM) network and a Light Gradient Boosting Machine (LightGBM). The model is designed to perform dual-stream learning: the LSTM module extracts medium- and long-term temporal patterns from historical TCI sequences, while the LightGBM module concurrently learns discriminative feature representations from the structured traffic data. A genetic algorithm (GA) is employed to optimize the fusion weights of the two components, constructing an adaptive and cohesive LightGBM-LSTM prediction model. The proposed framework was validated using real-world TCI data collected from three representative segments with varying congestion levels (mild, moderate, and severe) on Chengdu’s Third Ring Road, covering a period from September to October 2024. The experimental results demonstrate that the hybrid model significantly outperforms both standalone LSTM and LightGBM baselines across all test scenarios. Specifically, it achieved accuracy improvements of 4.87% and 33.06% in mildly congested sections, 26.80% and 22.32% in moderately congested sections, and 47.87% and 10.47% in severely congested sections, respectively, measured by the Mean Absolute Percentage Error (MAPE). These findings confirm that the proposed GA-optimized LightGBM-LSTM hybrid model effectively enhances TCI prediction precision and robustness by leveraging complementary strengths of sequence learning and feature engineering. The study provides a reliable and efficient analytical tool for short-term traffic state forecasting, offering valuable support for the development of data-driven and refined urban traffic management strategies.

References

1. Alafate, J., & Freund, Y. S. (2019). Faster boosting with smaller memory. Advances in Neural Information Processing Systems, 32. https://doi.org/10.48550/arXiv.1901.09047

2. Anitha, E. B., Aravinth, R., et al. (2019). Prediction of road traffic using naive bayes algorithm. Int. J. Eng. Res. Technol, 7(1), 1-4.

3. Cao, P., Dai, F., et al. (2021). A survey of traffic prediction based on deep neural network: Data, methods and challenges. Paper presented at the International conference on cloud computing. https://doi.org/10.1007/978-3-030-99191-3_2

4. Cheng, W., Li, J.-l., et al. (2022). Combination predicting model of traffic congestion index in weekdays based on LightGBM-GRU. Scientific reports, 12(1), 2912. https://doi.org/10.1038/s41598-022-06975-1

5. Chu, Z., Yu, J., et al. (2020). LPG-model: A novel model for throughput prediction in stream processing, using a light gradient boosting machine, incremental principal component analysis, and deep gated recurrent unit network. Information Sciences, 535, 107-129. https://doi.org/10.1016/j.ins.2020.05.042

6. Dissanayake, B., Hemachandra, O., et al. (2021). A comparison of ARIMAX, VAR and LSTM on multivariate short-term traffic volume forecasting. Paper presented at the Conference of open innovations association, FRUCT. https://10.1088/1757-899X/383/1/012043

7. Gu, Y., Lu, W., et al. (2019). An improved Bayesian combination model for short-term traffic prediction with deep learning. IEEE Transactions on Intelligent Transportation Systems, 21(3), 1332-1342. https://10.1109/TITS.2019.2939290

8. He, H., & Fan, Y. (2021). A novel hybrid ensemble model based on tree-based method and deep learning method for default prediction. Expert Systems with Applications, 176, 114899. https://doi.org/10.1016/j.eswa.2021.114899

9. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735

10. Kadiyala, A., Kumar, A., et al. (2018). Applications of python to evaluate the performance of decision tree‐based boosting algorithms. Environmental Progress, 37(2), 618-623. https://doi.org/10.1002/ep.12888

11. Kuang, L., Hua, C., et al. (2020). Traffic volume prediction based on multi-sources GPS trajectory data by temporal convolutional network. Mobile Networks Applications, 25(4), 1405-1417. https://10.1007/s11036-019-01458-6

12. Kumar, N., Martin, H., et al. (2024). Enhancing Deep Learning-Based City-Wide Traffic Prediction Pipelines Through Complexity Analysis. Data Science for Transportation, 6(3), 24. https://10.1007/s42421-024-00109-x

13. Li, F., Nie, W., et al. (2024). Network traffic prediction based on PSO-LightGBM-TM. Computer Networks, 254, 110810. https://10.1016/j.comnet.2024.110810

14. Li, L., Lin, H., et al. (2020). MF-TCPV: A machine learning and fuzzy comprehensive evaluation-based framework for traffic congestion prediction and visualization. Ieee Access, 8, 227113-227125. https://10.1109/ACCESS.2020.3043582

15. Li, N., Zou, F., et al. (2023). Vehicle traveling speed prediction based on LightGBM algorithm. Paper presented at the International Conference on Genetic and Evolutionary Computing. https://doi.org/10.1007/978-981-99-9412-0_1

16. Ma, X., Tao, Z., et al. (2015). Long short-term memory neural network for traffic speed prediction using remote microwave sensor data. Transportation Research Part C: Emerging Technologies, 54, 187-197. https://10.1016/j.trc.2015.03.014

17. Ministry of Public Security of the People’s Republic of China. (2025,01,18). https://www.mps.gov.cn /n2254314/6409334/c9939035/content.html.

18. Shi, X., Qi, H., et al. (2020). A spatial–temporal attention approach for traffic prediction. IEEE Transactions on Intelligent Transportation Systems, 22(8), 4909-4918. https://doi.org/10.1109/TITS.2020.2983651

19. Specifications for urban traffic performance evaluation. (2016). (Vol. 32). Beijing: Standards Press of China. GB/T33171-2016.

20. Yang, Z., Tang, R., et al. (2021). Short-term prediction of airway congestion index using machine learning methods. Transportation Research Part C: Emerging Technologies, 125, 103040. https://10.1016/j.trc.2021.103040

21. Zhang, X., Huang, K., et al. (2023). Urban short-term traffic flow prediction algorithm based on cnn-lstm model. Paper presented at the 2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE). https://doi.org/10.1109/ICCECE58074.2023.10135384

22. Zhao, S.-x., Wu, H.-w., et al. (2019). Traffic flow prediction based on optimized hidden Markov model. Paper presented at the Journal of Physics: Conference Series. https://doi.org/10.1088/1742-6596/1168/5/052001

23. Zhong, Y., Xie, X., et al. (2018). A new method for short-term traffic congestion forecasting based on LSTM. Paper presented at the IOP Conference Series: Materials Science and Engineering. https://10.1088/1757-899X/383/1/012043