Traffic speed sparse time series prediction model integrating spatiotemporal periodic features

Shan Jiang; Yuming Feng; Jiang Xiong; Shan Jiang; Yuming Feng; Jiang Xiong

doi:10.3934/math.2026570

AIMS Mathematics

2026, Volume 11, Issue 5: 13837-13864. doi: 10.3934/math.2026570

Previous Article Next Article

Research article

Traffic speed sparse time series prediction model integrating spatiotemporal periodic features

1.
School of Computer Science and Engineering, Chongqing Sanxia University of Science and Technology, Wanzhou, Chongqing 404100, China
2.
Key Laboratory of Intelligent Information Processing and Control, Chongqing Sanxia University of Science and Technology, Wanzhou, Chongqing 404100, China
3.
School of Mathematics and Statistics, Chongqing Sanxia University of Science and Technology, Wanzhou, Chongqing 404100, China

Received: 01 April 2026 Revised: 05 May 2026 Accepted: 11 May 2026 Published: 18 May 2026
MSC : 68T07

Sparse time series forecasting (SparseTSF) is a recently proposed lightweight multi-step forecasting model with advantages such as high computational efficiency and wide adaptability. However, the SparseTSF model also has some limitations, for example, it can only downsample a single main period, making it difficult to handle multi-period data. Based on the characteristics of traffic forecasting, this paper integrates weekly and spatial feature extraction modules to extract deep features that fuse daily, weekly, and spatial features. The SparseTSF model is then optimized to construct a multi-period spatial time series forecasting (MSTSF) model. This model is applied to traffic speed forecasting scenarios, aiming to reduce prediction error and achieve a balance between performance and parameter size. Experiments on the Guangzhou traffic and Performance Measurement System (PeMS) datasets show that the MSTSF model performs well in both prediction accuracy and model efficiency. Compared with mainstream deep learning models, this model achieves lower prediction errors. The MSTSF model has advantages such as small parameter size, high iteration efficiency, and fast inference speed, making it suitable for scenarios with limited computational resources. Our model achieves better results in traffic speed forecasting.
- Multi-step prediction,
- MSTSF,
- SparseTSF,
- periodic features,
- spatial features,
- traffic speed
Citation: Shan Jiang, Yuming Feng, Jiang Xiong. Traffic speed sparse time series prediction model integrating spatiotemporal periodic features[J]. AIMS Mathematics, 2026, 11(5): 13837-13864. doi: 10.3934/math.2026570

Related Papers:

Abstract

Sparse time series forecasting (SparseTSF) is a recently proposed lightweight multi-step forecasting model with advantages such as high computational efficiency and wide adaptability. However, the SparseTSF model also has some limitations, for example, it can only downsample a single main period, making it difficult to handle multi-period data. Based on the characteristics of traffic forecasting, this paper integrates weekly and spatial feature extraction modules to extract deep features that fuse daily, weekly, and spatial features. The SparseTSF model is then optimized to construct a multi-period spatial time series forecasting (MSTSF) model. This model is applied to traffic speed forecasting scenarios, aiming to reduce prediction error and achieve a balance between performance and parameter size. Experiments on the Guangzhou traffic and Performance Measurement System (PeMS) datasets show that the MSTSF model performs well in both prediction accuracy and model efficiency. Compared with mainstream deep learning models, this model achieves lower prediction errors. The MSTSF model has advantages such as small parameter size, high iteration efficiency, and fast inference speed, making it suitable for scenarios with limited computational resources. Our model achieves better results in traffic speed forecasting.

References

[1]	T. Alghamdi, K. Elgazzar, M. Bayoumi, T. Sharaf, S. Shah, Forecasting traffic congestion using ARIMA modeling, In: 15th International Wireless Communications & Mobile Computing Conference (IWCMC), 2019, 24–28. https://doi.org/10.1109/IWCMC.2019.8766698
[2]	B. Dissanayake, O. Hemachandra, N. Lakshitha, D. Haputhanthri, A. Wijayasiri, A comparison of ARIMAX, VAR and LSTM on multivariate short-term traffic volume forecasting, In: Conference of Open Innovations Association, FRUCT, 2021,564–570.
[3]	M. Castro-Neto, Y. S. Jeong, M. K. Jeong, L. D. Han, Online-SVR for short-term traffic flow prediction under typical and atypical traffic conditions, Expert Syst. Appl., 36 (2009), 6164–6173. https://doi.org/10.1016/j.eswa.2008.07.069 doi: 10.1016/j.eswa.2008.07.069
[4]	R. G. Purnima, S. Kiran, P. A. RG, M. S. Vinotheni, P. Ramesh, Traffic light control system and traffic prediction using machine learning with SVR and RFR, In: 2024 4th International Conference on Mobile Networks and Wireless Communications (ICMNWC), 2024. https://doi.org/10.1109/ICMNWC63764.2024.10872340
[5]	X. Luo, D. Li, Y. Yang, S. Zhang, Spatiotemporal traffic flow prediction with KNN and LSTM, J. Adv. Transport., 2019 (2019), 4145353. https://doi.org/10.1155/2019/4145353 doi: 10.1155/2019/4145353
[6]	Y. Kong, Z. Wang, Y. Nie, T. Zhou, S. Zohren, Y. Liang, et al., Unleashing the power of LSTM in long-term time series forecasting, In: Proceedings of the AAAI Conference on Artificial Intelligence, 39 (2025), 11968–11976. https://doi.org/10.1609/aaai.v39i11.33303
[7]	S. Yan, F. Wen, J. Wu, Capacity matching method for mountain scenic areas during holidays based on highway traffic data, in Chinese, J. Jilin Uni., 55 (2025), 1576–1587.
[8]	J. Fan, K. Zhang, Y. Huang, Y. Zhu, B. Chen, Parallel spatio-temporal attention-based TCN for multivariate time series prediction, Neural Comput. Applic., 35 (2023), 13109–13118. https://doi.org/10.1007/s00521-021-05958-z doi: 10.1007/s00521-021-05958-z
[9]	Z. Dou, D. Guo, DPSTCN: Dynamic pattern-aware spatio-temporal convolutional networks for traffic flow forecasting, ISPRS Int. J. Geo-Inform., 14 (2024), 10. https://doi.org/10.3390/ijgi14010010 doi: 10.3390/ijgi14010010
[10]	A. Zeng, M. Chen, L. Zhang, Q. Xu, Are transformers effective for time series forecasting? In: Proceedings of the AAAI Conference on Artificial Intelligence, 37 (2023), 11121–11128. https://doi.org/10.1609/aaai.v37i9.26317
[11]	Y. Song, R. Luo, T. Zhou, C. Zhou, R. Su, Graph attention informer for long-term traffic flow prediction under the impact of sports events, Sensors, 24 (2024), 4796. https://doi.org/10.3390/s24154796 doi: 10.3390/s24154796
[12]	S. Liu, X. Wang, An improved transformer based traffic flow prediction model, Sci. Rep., 15 (2025), 8284. https://doi.org/10.1038/s41598-025-92425-7 doi: 10.1038/s41598-025-92425-7
[13]	X. Zhang, Z. Zhao, J. Li, ARDE-N-BEATS: An evolutionary deep learning framework for urban traffic flow prediction, IEEE Int. Things J., 10 (2022), 2391–2403. https://doi.org/10.1109/JIOT.2022.3212056 doi: 10.1109/JIOT.2022.3212056
[14]	D. Cheng, Y. Zhang, H. Li, EEMD-TiDE-based passenger flow prediction for urban rail transit, Electronics, 15 (2026), 529. https://doi.org/10.3390/electronics15030529 doi: 10.3390/electronics15030529
[15]	T. Zuo, S. Tang, L. Zhang, H. Kang, H. Song, P. Li, An enhanced TimesNet-SARIMA model for predicting outbound subway passenger flow with decomposition techniques, Appl. Sci., 15 (2025), 2874. https://doi.org/10.3390/app15062874 doi: 10.3390/app15062874
[16]	I. E. Livieris, C-kan: A new approach for integrating convolutional layers with kolmogorov–arnold networks for time-series forecasting, Mathematics, 12 (2024), 3022. https://doi.org/10.3390/math12193022 doi: 10.3390/math12193022
[17]	L. A. I. Sicong, H. U. Yuehong. STKAN: Kolmogorov-arnold networks for spatio-temporal forecasting, In: ICLR 2026 Conference, 2016.
[18]	G. Jin, S. Lai, X. Hao, J. Zhang, M. Zhang, M3-net: A cost-effective graph-free mlp-based model for traffic prediction, In: Proceedings of the 34th ACM International Conference on Information and Knowledge Management, 2025, 4847–4851. https://doi.org/10.1145/3746252.3760815
[19]	W. Duan, H. Rao, W. Huang, X. He, Minimalist traffic prediction: Linear layer is all you need, preprint paper, 2023. https://doi.org/10.48550/arXiv.2308.10276
[20]	D. Wang, G. Guo, T. Ouyang, D. Yu, H. Zhang, B. Li, et al., A lightweight spatio-temporal neural network with sampling-based time series decomposition for traffic forecasting, IEEE Trans. Intell. Transport. Syst., 26 (2025), 8682–8693. https://doi.org/10.1109/TITS.2025.3552010 doi: 10.1109/TITS.2025.3552010
[21]	Q. Zheng, M. Shao, Y. Zhang, TLAST: A time-lag aware spatial-temporal transformer for traffic flow forecasting, IEEE Trans. Intell. Transport. Syst., 26 (2025), 13144–13158. https://doi.org/10.1109/TITS.2025.3583391 doi: 10.1109/TITS.2025.3583391
[22]	Z. Zhong, H.Wu, PASTNet: A prototype-pattern-driven lightweight model for traffic flow forecasting, J. Yunnan Uni., 2025. https://doi.org/10.7540/j.ynu.20250193 doi: 10.7540/j.ynu.20250193
[23]	J. Zhao, F. Zhuo, Q. Sun, Q. Li, Y. Hua, J. Zhao, DSFormer-LRTC: Dynamic spatial transformer for traffic forecasting with low-rank tensor compression, IEEE Trans. Intell. Transport. Syst., 25 (2024), 16323–16335. https://doi.org/10.1109/TITS.2024.3436523 doi: 10.1109/TITS.2024.3436523
[24]	J. Xu, J. Yang, Y. Huang, L. Y. Por, X. Chen, C. Zhao, DFGNet: A dual-pathway graph neural network via frequency decomposition for spatiotemporal forecasting, Expert Syst. Appl., 297 (2026), 129518. https://doi.org/10.1016/j.eswa.2025.129518 doi: 10.1016/j.eswa.2025.129518
[25]	J. Chen, Q. Shao, D. Chen, W. Yu, Decoupling spatio-temporal prediction: When lightweight large models meet adaptive Hhypergraphs, In: Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2025,167–178. https://doi.org/10.1145/3711896.3736904
[26]	S. Lin, W. Lin, W. Wu, H. Chen, J. Yang, Sparsetsf: Modeling long-term time series forecasting with 1k parameters, preprint paper, 2024. https://doi.org/10.48550/arXiv.2405.00946
[27]	M. Reyad, A. M. Sarhan, M. Arafa, A modified Adam algorithm for deep neural network optimization, Neural Comput. Applic., 35 (2023), 17095–17112. https://doi.org/10.1007/s00521-023-08568-z doi: 10.1007/s00521-023-08568-z
[28]	OpenData V12.0-Large-scale Traffic Speed Data Set, OpenITS Org, 2021. Available from: http://www.openits.cn/openData4/824.jhtml. Accessed: 2025-05-09.
[29]	California department of transportation, PeMS, 2025. Available from: http://pems.dot.ca.gov.
[30]	S. Somvanshi, S. A. Javed, M. M. Islam, D. Pandit, S. Das, A survey on kolmogorov-arnold network, ACM Comput. Surveys, 58 (2025), 1–35. https://doi.org/10.1145/3743128 doi: 10.1145/3743128
[31]	D. Salinas, V. Flunkert, J. Gasthaus, T. Januschowski, DeepAR: Probabilistic forecasting with autoregressive recurrent networks, Int. J. Forec., 36 (2020), 1181–1191. https://doi.org/10.1016/j.ijforecast.2019.07.001 doi: 10.1016/j.ijforecast.2019.07.001
[32]	T. Zhou, Z. Ma, Q. Wen, X. Wang, L. Sun, R. Jin, Fedformer: Frequency enhanced decomposed transformer for long-term series forecasting, In: Proceedings of the 39th International Conference on Machine Learning, PMLR, 162 (2022), 27268–27286.
[33]	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: Proceedings of the AAAI Conference on Artificial Intelligence, 35 (2021), 11106–11115. https://doi.org/10.1609/aaai.v35i12.17325
[34]	M. Kasprzyk, P. Pełka, B. N. Oreshkin, G. Dudek, Enhanced N-BEATS for Mid-Term electricity demand forecasting, Appl. Soft Comput., 182 (2025), 113575. https://doi.org/10.1016/j.asoc.2025.113575 doi: 10.1016/j.asoc.2025.113575
[35]	C. Challu, K. G. Olivares, B. N. Oreshkin, F. G. Ramirez, M. M. Canseco, A Dubrawski Nhits: Neural hierarchical interpolation for time series forecasting, In: Proceedings of the AAAI Conference on Artificial Intelligence, 37 (2023), 6989–6997. https://doi.org/10.1609/aaai.v37i6.25854
[36]	H. Zheng, Y. Lu, Z. Sun, J. Panneerselvam, X. Sun, L. Liu, Energy optimisation in cloud datacentres with MC-TIDE: Mixed channel time-series dense encoder for workload forecasting, Appl. Energy, 374 (2024), 123903. https://doi.org/10.1016/j.apenergy.2024.123903129518 doi: 10.1016/j.apenergy.2024.123903129518
[37]	X. Huang, J. Tang, Y. Shen, Long time series of ocean wave prediction based on PatchTST model, Ocean Eng., 301 (2024), 117572. https://doi.org/10.1016/j.oceaneng.2024.117572 doi: 10.1016/j.oceaneng.2024.117572
[38]	H. Wu, J. Xu, J. Wang, M. Long, Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting, Adv. Neur. Inform. Proc. Syst., 34 (2021), 22419–22430.
[39]	A. Bhatnagar, A. Paule, T. Schuermann, S. Reiter, O. Bringmann, On-device adaptive battery power prediction for electric vehicles, In: 2025 IEEE 11th International Conference on Edge Computing and Scalable Cloud, IEEE, 2025. https://doi.org/10.1109/EdgeCom66327.2025.00026
[40]	H. Li, H. Lv, Q. Liu, X. Liu, Y. Zhang, X. Zhou, A contextually enhanced self-attention dilated RNN for load forecasting, In: 2025 IEEE Smart World Congress (SWC). IEEE, 2025. https://doi.org/10.1109/SWC65939.2025.00110

Reader Comments

Your name:*

Email:*
© 2026 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)