A variational mode decomposition assisted temporal-frequency pure convolutional neural network for highway traffic flow forecasting

Yifei Zheng; Xiang Wang; Xintong Liu; Shaoweihua Liu; Zhiyuan Liu; Kai Huang; Yifei Zheng; Xiang Wang; Xintong Liu; Shaoweihua Liu; Zhiyuan Liu; Kai Huang

doi:10.3934/era.2025320

Electronic Research Archive

2025, Volume 33, Issue 11: 7247-7276. doi: 10.3934/era.2025320

Previous Article Next Article

Research article Special Issues

A variational mode decomposition assisted temporal-frequency pure convolutional neural network for highway traffic flow forecasting

1.
School of Cyber Science and Engineering, Southeast University, Nanjing, China
2.
School of Instrument Science and Engineering, State Key Laboratory of comprehensive PNT Network and Equipment Technology, Southeast University, Nanjing, China
3.
Zhejiang Communications Investment Group Technology Research Global Co., Ltd., Hangzhou, China
4.
Jiangsu Key Laboratory of Urban ITS, Jiangsu Province Collaborative Innovation Center of Modern Urban Traffic Technologies, School of Transportation, Southeast University, Nanjing, China

Received: 29 July 2025 Revised: 01 November 2025 Accepted: 11 November 2025 Published: 28 November 2025

This paper addressed the problem of highway traffic flow multivariate time series forecasting with the challenges of variate heterogeneity. To improve forecast performance without a heavy computational burden in large-scale networks, we innovatively introduced variational mode decomposition and established a decomposition-assisted multi-tasking deep learning forecasting architecture. To improve variate-specific pattern learning and mitigate pattern mixing, we proposed a novel temporal-frequency pure convolutional neural network incorporating discrete Fourier transform, deepwise convolution, and batchwise feedforward neural network. To verify the proposed model, we conducted a case study on a regional network located in Jiangsu, China. Results demonstrate strong forecast performance and efficient computation. The proposed model offers suitability toward highway operators for large-scale engineering deployment and better facilitates their managerial actions.
- multivariate time series forecasting,
- deep learning,
- signal decomposition,
- frequency learning,
- convolutional neural network
Citation: Yifei Zheng, Xiang Wang, Xintong Liu, Shaoweihua Liu, Zhiyuan Liu, Kai Huang. A variational mode decomposition assisted temporal-frequency pure convolutional neural network for highway traffic flow forecasting[J]. Electronic Research Archive, 2025, 33(11): 7247-7276. doi: 10.3934/era.2025320

Related Papers:

Abstract

This paper addressed the problem of highway traffic flow multivariate time series forecasting with the challenges of variate heterogeneity. To improve forecast performance without a heavy computational burden in large-scale networks, we innovatively introduced variational mode decomposition and established a decomposition-assisted multi-tasking deep learning forecasting architecture. To improve variate-specific pattern learning and mitigate pattern mixing, we proposed a novel temporal-frequency pure convolutional neural network incorporating discrete Fourier transform, deepwise convolution, and batchwise feedforward neural network. To verify the proposed model, we conducted a case study on a regional network located in Jiangsu, China. Results demonstrate strong forecast performance and efficient computation. The proposed model offers suitability toward highway operators for large-scale engineering deployment and better facilitates their managerial actions.

References

[1]	K. Huang, Z. Zhang, X. Wang, Y. Tao, Z. Liu, Life-cycle carbon emissions of autonomous electric vehicles in varying traffic situations, Transp. Res. Part D Transp. Environ. , 146 (2025), 104871. https://doi.org/10.1016/j.trd.2025.104871 doi: 10.1016/j.trd.2025.104871
[2]	X. Zhang, R. R. Chowdhury, R. K. Gupta, J. Shang, Large language models for time series: A survey, in the 33rd International Joint Conference on Artificial Intelligence (IJCAI), (2024), 8335–8343. https://doi.org/10.24963/ijcai.2024/921
[3]	Z. Liu, Z. Zhou, Z. Gu, S. Liu, P. Liu, Y. Zhang, et al., TRIP: Transport reasoning with intelligence progression — a foundation framework, Transp. Res. Part C Emerging Technol. , 179 (2025), 105260. https://doi.org/10.1016/j.trc.2025.105260 doi: 10.1016/j.trc.2025.105260
[4]	B. M. Williams, Multivariate vehicular traffic flow prediction: Evaluation of arimax modeling, Transp. Res. Rec. , 1776 (2001), 194–200. https://doi.org/10.3141/1776-25 doi: 10.3141/1776-25
[5]	N. G. Polson, V. O. Sokolov, Deep learning for short-term traffic flow prediction, Transp. Res. Part C Emerging Technol. , 79 (2017), 1–17. https://doi.org/10.1016/j.trc.2017.02.024 doi: 10.1016/j.trc.2017.02.024
[6]	Z. Liu, C. Lyu, Z. Wang, S. Wang, P. Liu, Q. Meng, A Gaussian-process-based data-driven traffic flow model and its application in road capacity analysis, IEEE Trans. Intell. Transp. Syst. , 24 (2023), 1544–1563. https://doi.org/10.1109/TITS.2022.3223982 doi: 10.1109/TITS.2022.3223982
[7]	Z. Liu, C. Lyu, J. Huo, S. Wang, J. Chen, Gaussian process regression for transportation system estimation and prediction problems: The deformation and a Hat Kernel, IEEE Trans. Intell. Transp. Syst. , 23 (2022), 22331–22342. https://doi.org/10.1109/TITS.2022.3155527 doi: 10.1109/TITS.2022.3155527
[8]	M. Liu, A. Zeng, Q. Lai, Q. Xu, T-WaveNet: Tree-structured wavelet neural network for sensor-based time series analysis, preprint, arXiv: 2012.05456.
[9]	Z. Zhao, W. Chen, X. Wu, P. C. Y. Chen, J. Liu, LSTM network: A deep learning approach for short-term traffic forecast, IET Intell. Transp. Syst. , 11 (2017), 68–75. https://doi.org/10.1049/iet-its.2016.0208 doi: 10.1049/iet-its.2016.0208
[10]	A. Zeng, M. Chen, L. Zhang, Q. Xu, Are transformers effective for time series forecasting?, in the 37th AAAI Conference on Artificial Intelligence (AAAI), AAAI Press, (2023), 11121–11128. https://doi.org/10.1609/aaai.v37i9.26317
[11]	A. Das, W. Kong, A. Leach, S. Mathur, R. Sen, R. Yu, Long-term forecasting with tide: Time-series dense encoder, preprint, arXiv: 2304.08424.
[12]	S. Bai, J. Z. Kolter, V. Koltun, An empirical evaluation of generic convolutional and recurrent networks for sequence modeling, preprint, arXiv: 1803.01271.
[13]	B. Yu, H. Yin, Z. Zhu, Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting, in the 27th International Joint Conference on Artificial Intelligence (IJCAI), (2018), 3634–3640, https://doi.org/10.24963/ijcai.2018/505
[14]	Z. Wu, S. Pan, G. Long, J. Jiang, C. Zhang, Graph wavenet for deep spatial-temporal graph modeling, in the 28th International Joint Conference on Artificial Intelligence (IJACI), (2019), 1907–1913.
[15]	Z. Wu, S. Pan, G. Long, J. Jiang, X. Chang, C. Zhang, Connecting the Dots: Multivariate time series forecasting with graph neural networks, in the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, (2020), 753–763. https://doi.org/10.1145/3394486.3403118
[16]	Z. Gong, Y. Tang, J. Liang, PatchMixer: A patch-mixing architecture for long-term time series forecasting, preprint, arXiv: 2310.00655.
[17]	Y. Nie, N. H. Nguyen, P. Sinthong, J. Kalagnanam, A time series is worth 64 words: Long-term forecasting with transformers, in the 11th International Conference on Learning Representations (ICLR), (2023), 1–24.
[18]	D. Luo, X. Wang, ModernTCN: A modern pure convolution structure for general time series analysis, in the 12th International Conference on Learning Representations (ICLR), (2024), 1–43.
[19]	A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, et al., Attention is all you need, in Advances in Neural Information Processing Systems 30 (NeurlPS), (2017), 1–11.
[20]	Y. Zhang, J. Yan, Crossformer: Transformer utilizing cross-dimension dependency for multivariate time series forecasting, in the 11th International Conference on Learning Representations (ICLR), (2023), 1–21.
[21]	H. Wu, J. Xu, J. Wang, M. Long, Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting, in Advances in Neural Information Processing Systems 34 (NeurlPS), (2021), 22419–22430.
[22]	H. Zhou, S. Zhang, J. Peng, S. Zhang, J. Li, H. Xiong, et al., Informer: Beyond efficient transformer for long sequence time-series forecasting, in the 35th AAAI Conference on Artificial Intelligence (AAAI), 35 (2021), 11106–11115. https://doi.org/10.1609/aaai.v35i12.17325
[23]	S. Guo, Y. Lin, N. Feng, C. Song, H. Wan, Atention based spatial-temporal graph convolutional networks for traffic flow forecasting, in the 33th AAAI Conference on Artificial Intelligence (AAAI), 33 (2019), 922–929. https://doi.org/10.1609/aaai.v33i01.3301922
[24]	Q. Wu, C. Yang, W. Zhao, Y. He, D. Wipf, J. Yan, DIFFormer: Scalable (graph) transformers induced by energy constrained diffusion, in the 11th International Conference on Learning Representations (ICLR), (2023), 1–26.
[25]	Y. Liu, T. Hu, H. Zhang, H. Wu, S. Wang, L. Ma, et al., iTransformer: Inverted transformers are effective for time series forecasting, in the 12th International Conference on Learning Representations (ICLR), (2024), 1–25.
[26]	A. Katharopoulos, A. Vyas, N. Pappas, F. Fleuret, Transformers are RNNs: Fast autoregressive transformers with linear attention, in the 37th International Conference on Machine Learning (ICML), (2020), 5156–5165.
[27]	C. Ying, T. Cai, S. Luo, S. Zheng, G. Ke, D. He, et al., Do transformers really perform badly for graph representation?, in Advances in Neural Information Processing Systems 34 (NeurlPS), (2021), 28877–28888.
[28]	H. Wu, T. Hu, Y. Liu, H. Zhou, J. Wang, M. Long, TimesNet: Temporal 2d-variation modeling for general time series analysis, in the 11th International Conference on Learning Representations (ICLR), (2023), 1–23.
[29]	Z. Xu, A. Zeng, Q. Xu, FITS: Modeling time series with 10k parameters, in the 12th International Conference on Learning Representations (ICLR), (2024), 1–24.
[30]	H. Wang, L. Pan, Z. Chen, D. Yang, S. Zhang, Y. Yang, et al., FreDF: Learning to forecast in the frequency domain, in the 13th International Conference on Learning Representations (ICLR), (2025), 1–30.
[31]	Y. He, L. Li, X. Zhu, K. L. Tsui, Multi-graph convolutional-recurrent neural network (MGC-RNN) for short-term forecasting of transit passenger flow, IEEE Trans. Intell. Transp. Syst. , 23 (2022), 18155–18174. https://doi.org/10.1109/TITS.2022.3150600 doi: 10.1109/TITS.2022.3150600
[32]	C. Zheng, X. Fan, C. Wen, L. Chen, C. Wang, J. Li, DeepSTD: Mining spatio-temporal disturbances of multiple context factors for citywide traffic flow prediction, IEEE Trans. Intell. Transp. Syst. , 21 (2019), 3744–3755. https://doi.org/10.1109/TITS.2019.2932785 doi: 10.1109/TITS.2019.2932785
[33]	J. Lu, X. Han, Y. Sun, S. Yang, CATS: Enhancing multivariate time series forecasting by constructing auxiliary time series as exogenous variables, in the 41st International Conference on Machine Learning (ICML), (2024), 1–17.
[34]	X. Huang, S. Belongie, Arbitrary style transfer in real-time with adaptive instance normalization, in 2017 IEEE International Conference on Computer Vision (ICCV), IEEE, (2017), 1510–1519. https://doi.org/10.1109/ICCV.2017.167
[35]	N. Bjorck, C. P. Gomes, B. Selman, K. Q. Weinberger, Understanding batch normalization, in Advances in Neural Information Processing Systems 31 (NeurlPS), (2018), 1–12.
[36]	T. Kim, J. Kim, Y. Tae, C. Park, J. H. Choi, J. Choo, Reversible instance normalization for accurate time-series forecasting against distribution shift, in the 10th International Conference on Learning Representations (ICLR), (2022), 1–25.
[37]	K. Dragomiretskiy, D. Zosso, Variational mode decomposition, IEEE Trans. Signal Process. , 62 (2013), 531–544. https://doi.org/10.1109/TSP.2013.2288675. doi: 10.1109/TSP.2013.2288675
[38]	A. Taha, N. Nazih, P. Makeen, Wind speed prediction based on variational mode decomposition and advanced machine learning models in zaafarana, Egypt, Sci. Rep. , 15 (2025), 15599. https://doi.org/10.1038/s41598-025-98543-6 doi: 10.1038/s41598-025-98543-6
[39]	E. A. Tuncar, S. Saglam, B. Oral, A review of short-term wind power generation forecasting methods in recent technological trends, Energy Rep. , 12 (2024), 197–209. https://doi.org/10.1016/j.egyr.2024.06.006 doi: 10.1016/j.egyr.2024.06.006
[40]	G. Liu, Y. Ma, N. Wang, Rolling bearing fault diagnosis based on SABO-VMD and weighted Manhattan-KNN, Sensors, 24 (2024), 5003. https://doi.org/10.3390/s24155003 doi: 10.3390/s24155003
[41]	N. ur Rehman, H. Aftab, Multivariate variational mode decomposition, IEEE Trans. Signal Process. , 67 (2019), 6039–6052. https://doi.org/10.1109/TSP.2019.2951223 doi: 10.1109/TSP.2019.2951223
[42]	C. Xu, J. Yang, T. Zhang, K. Li, K. Zhang, Adaptive parameter selection variational mode decomposition based on a novel hybrid entropy and its applications in locomotive bearing diagnosis, Measurement, 217 (2023), 113110. https://doi.org/10.1016/j.measurement.2023.113110 doi: 10.1016/j.measurement.2023.113110
[43]	H. Jia, P. Cao, T. Liang, C. F. Caiafa, Z. Sun, Y. Kushihashi, et al., Short-time variational mode decomposition, Signal Process. , 238 (2026), 110203. https://doi.org/10.1016/j.sigpro.2025.110203 doi: 10.1016/j.sigpro.2025.110203
[44]	K. Dragomiretskiy, D. Zosso, Two-dimensional variational mode decomposition, in Energy Minimization Methods in Computer Vision and Pattern Recognition, Springer, (2015), 197–208. https://doi.org/10.1007/978-3-319-14612-6_15
[45]	D. P. Bertsekas, Chapter 1 - Introduction, in Constrained Optimization and Lagrange Multiplier Methods, Academic Press, (1982), 1–94, https://doi.org/10.1016/B978-0-12-093480-5.50005-2
[46]	Z. Liu, Y. Wang, Q. Cheng, H. Yang, Analysis of the information entropy on traffic flows, IEEE Trans. Intell. Transp. Syst. , 23 (2022), 18012–18023. https://doi.org/10.1109/TITS.2022.3155933 doi: 10.1109/TITS.2022.3155933
[47]	H. Liu, Z. Dong, R. Jiang, J. Deng, J. Deng, Q. Chen, et al., Spatio-Temporal adaptive embedding makes vanilla transformer SOTA for traffic forecasting, in the 32nd ACM International Conference on Information and Knowledge Management (CIKM), (2023), 4125–4129. https://doi.org/10.1145/3583780.3615160

Reader Comments

Your name:*

Email:*
© 2025 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)