Traffic flow prediction driven by heterogeneity decoupling and feature layered modeling

doi:10.3785/j.issn.1008-973X.2026.06.005

Journal of ZheJiang University (Engineering Science)

2026, Vol. 60

Issue (6): 1185-1195 DOI: 10.3785/j.issn.1008-973X.2026.06.005

Traffic flow prediction driven by heterogeneity decoupling and feature layered modeling

Yue HOU(

),Jinlong XIE,Lindong ZHANG,Jie YIN,Tiantian WANG

School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China

Download:

HTML

PDF(1709KB) HTML
Export: BibTeX | EndNote (RIS)

Abstract

A new traffic flow prediction model named CFHD-Former was proposed in order to address the limitations of existing traffic flow prediction models that struggle to capture dynamic variations across different time slices and cannot adequately consider the heterogeneous characteristics of traffic volume distributions across regions. An adaptive high-frequency heterogeneity module and a progressive optimization mechanism were introduced to enhance its adaptability to traffic states under different time slices. A core flow node identification module was employed building on the captured temporal heterogeneity feature in order to partition the road network into core and non-core flow networks based on nodal traffic flow characteristics. Heterogeneous modeling of the two types of road network was implemented via a spatial encoder. A frequency-domain autocorrelation MAE loss function was incorporated during backpropagation in order to consider the dependencies among different time steps within the prediction sequence, thereby reducing multi-step prediction errors. The experimental results demonstrated that the MAE of the proposed CFHD-Former model was reduced by 1.70%, 4.58% and 4.44% on the PEMS04, PEMS08 and METR-LA datasets, respectively compared with the best-performing baseline model. Results verified the effectiveness of CFHD-Former in modeling the spatio-temporal heterogeneity of complex road networks and provided a new solution for urban traffic flow prediction.

Key words： traffic flow prediction spatio-temporal heterogeneity core flow point identification heterogeneous modeling heterogeneous characteristic

Received: 06 March 2025 Published: 06 May 2026

CLC:

U 491

Fund: 国家自然科学基金资助项目（62063014，62363020）.

	Service
	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Yue HOU
	Jinlong XIE
	Lindong ZHANG
	Jie YIN
	Tiantian WANG

Cite this article:

Yue HOU,Jinlong XIE,Lindong ZHANG,Jie YIN,Tiantian WANG. Traffic flow prediction driven by heterogeneity decoupling and feature layered modeling. Journal of ZheJiang University (Engineering Science), 2026, 60(6): 1185-1195.

URL:

https://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2026.06.005 OR https://www.zjujournals.com/eng/Y2026/V60/I6/1185

异质性解耦与特征分层建模驱动的交通流预测

为了解决现有交通流预测模型难以捕捉不同时间片下的动态变化，且无法充分考虑不同区域流量分布的差异化特征的问题，提出新型的交通流预测模型CFHD-Former. 该模型引入自适应高频异质化模块和渐进调优机制，增强对不同时间片交通状态的适应性. 在捕捉时间异质化特征的基础上，利用核心流动点识别模块，根据节点流量特征来划分核心与非核心流动路网，通过空间编码器实现对2类路网的异质化建模. 在反向传播过程中引入频域自相关MAE损失函数，考虑预测序列不同时间步间的依赖关系，达到降低多步预测误差的目的. 实验结果表明，相较于最优的基线模型，在PEMS04、PEMS08及METRLA 3个数据集上，所提CFHD-Former模型的MAE分别降低了1.70%、4.58%、4.44%. 结果验证了CFHD-Former模型在复杂路网时空异质性建模方面的有效性，为城市交通流预测提供了新的解决方案.

关键词： 交通流预测, 时空异质性, 核心流动点识别, 异质化建模, 差异化特征

Fig.1 Residential area and downtown distribution

Fig.2 Comparison of time-varying flow characteristic between downtown and residential area

Fig.3 Autocorrelation analysis of input-prediction sequence

Fig.4 Architecture diagram of CFHD-Former model

Fig.5 Architecture of temporal Transformer

Fig.6 Schematic diagram of node inflow and outflow capacity calculation

Fig.7 Architecture of spatial Transformer

Fig.8 Comparison of recursive and direct prediction paradigm

Fig.9 Diagram of frequency domain transformation

Tab.1 Basic information of Caltrans PeMS dataset

Tab.2 Comparison of prediction accuracy of different baseline model based on PEMS04, PEMS08 and METRLA dataset

Fig.10 Analysis of predictive performance of different ablation model

Tab.3 Predictive performance of ablation model at different time step

Fig.11 Sensitivity analysis of key hyperparameter

Fig.12 Geospatial distribution of core flow point

Fig.13 Comparison of flow time-varying pattern between core and non-core flow point

Fig.14 Dimensionality reduction visualization of t-SNE based on data from different date

Fig.15 Comparison of CFHD-Former and DGCRN fitting true value　　　

Fig.16 Comparison of prediction error between CFHD-Former and baseline model under different perturbation ratio noise


[1]	SUN Y, ZHANG G, YIN H Passenger flow prediction of subway transfer stations based on nonparametric regression model[J]. Discrete Dynamics in Nature and Society, 2014, 2014 (1): 397154

[2]	TONG J, GU X, ZHANG M, et al. Traffic flow prediction based on improved SVR for VANET [C]//Proceedings of the 4th International Conference on Advanced Electronic Materials, Computers and Software Engineering. Changsha: IEEE, 2021: 402–405.

[3]	ZAREMBA W, SUTSKEVER I, VINYALS O. Recurrent neural network regularization [EB/OL]. (2015-02-19). https://arxiv.org/abs/1409.2329.

[4]	GRAVES A. Long short-term memory [M]//Supervised sequence labelling with recurrent neural networks. Berlin: Springer, 2012: 37–45.

[5]	HAO S, LEE D H, ZHAO D Sequence to sequence learning with attention mechanism for short-term passenger flow prediction in large-scale metro system[J]. Transportation Research Part C: Emerging Technologies, 2019, 107: 287- 300 doi: 10.1016/j.trc.2019.08.005

[6]	KIPF T N, WELLING M. Semi-supervised classification with graph convolutional networks [EB/OL]. (2019-06-15). https://arxiv.org/abs/1609.02907.

[7]	LI Z, XIONG G, CHEN Y, et al. A hybrid deep learning approach with GCN and LSTM for traffic flow prediction [C]//Proceedings of the IEEE Intelligent Transportation Systems Conference. Auckland: IEEE, 2019: 1929-1933.

[8]	KONG W, GUO Z, LIU Y Spatio-temporal pivotal graph neural networks for traffic flow forecasting[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2024, 38 (8): 8627- 8635 doi: 10.1609/aaai.v38i8.28707

[9]	SHAO Z, ZHANG Z, WANG F, et al. Spatial-temporal identity: a simple yet effective baseline for multivariate time series forecasting [C]//Proceedings of the 31st ACM International Conference on Information and Knowledge Management. Atlanta: ACM, 2022: 4454-4458.

[10]	DONG Z, JIANG R, GAO H, et al. Heterogeneity-informed meta-parameter learning for spatiotemporal time series forecasting [C]//Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. Barcelona: ACM, 2024: 631-641.

[11]	PHOLSENA K, PAN L, ZHENG Z Mode decomposition based deep learning model for multi-section traffic prediction[J]. World Wide Web, 2020, 23 (4): 2513- 2527 doi: 10.1007/s11280-020-00791-1

[12]	邵春福. 交通规划原理 [M]. 2版. 北京: 中国铁道出版社, 2014: 96–97.

[13]	LIU S, GHOSH R, MOTANI M. Towards better long-range time series forecasting using generative forecasting [EB/OL]. (2022-08-05). https://arxiv.org/abs/2212.06142.

[14]	WANG H, PAN L, CHEN Z, et al. FreDF: learning to forecast in the frequency domain [EB/OL]. (2024-02-04). https://arxiv.org/abs/2402.02399.

[15]	WILLIAMS B M, HOEL L A Modeling and forecasting vehicular traffic flow as a seasonal ARIMA process: theoretical basis and empirical results[J]. Journal of Transportation Engineering, 2003, 129 (6): 664- 672 doi: 10.1061/(ASCE)0733-947X(2003)129:6(664)

[16]	DENG J, CHEN X, JIANG R, et al. ST-norm: spatial and temporal normalization for multi-variate time series forecasting [C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. [S. l. ]: ACM, 2021: 269-278.

[17]	LIU M, ZENG A, CHEN M, et al SCINet: Time series modeling and forecasting with sample convolution and interaction[J]. Advances in Neural Information Processing Systems, 2022, 35: 5816- 5828

[18]	YU B, YIN H, ZHU Z. Spatio-temporal graph convolutional networks: a deep learning framework for traffic forecasting [EB/OL]. (2018-07-12). https://arxiv.org/abs/1709.04875.

[19]	WU Z, PAN S, LONG G, et al. Graph WaveNet for deep spatial-temporal graph modeling [EB/OL]. (2019-05-31). https://arxiv.org/abs/1906.00121.

[20]	LI Y, YU R, SHAHABI C, et al. Diffusion convolutional recurrent neural network: data-driven traffic forecasting [EB/OL]. (2018-02-22). https://arxiv.org/abs/1707.01926.

[21]	LI F, FENG J, YAN H, et al Dynamic graph convolutional recurrent network for traffic prediction: benchmark and solution[J]. ACM Transactions on Knowledge Discovery from Data, 2023, 17 (1): 1- 21

[22]	ZHENG C, FAN X, WANG C, et al GMAN: a graph multi-attention network for traffic prediction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34 (1): 1234- 1241 doi: 10.1609/aaai.v34i01.5477

[23]	DUAN W, HE X, ZHOU Z, et al. Localised adaptive spatial-temporal graph neural network [C]//Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. Long Beach: ACM, 2023: 448-458.

[1]	Yue HOU,Tiantian WANG,Xin ZHANG,Jie YIN. Traffic flow forecasting with multi-resolution trend period decoupling interaction[J]. Journal of ZheJiang University (Engineering Science), 2025, 59(7): 1362-1372.

[2]	Xinjian XIANG,Tianshun YUAN,Yaqiang HE,Chengli WANG. Traffic flow prediction based on time series decomposition and soft thresholding temporal convolution[J]. Journal of ZheJiang University (Engineering Science), 2025, 59(7): 1353-1361.

[3]	Li MA,Yongshun WANG,Yao HU,Lei FAN. Pre-trained long-short spatiotemporal interleaved Transformer for traffic flow prediction applications[J]. Journal of ZheJiang University (Engineering Science), 2025, 59(4): 669-678.

[4]	Yuxuan LIU,Yizhi LIU,Zhuhua LIAO,Zhengbiao ZOU,Jingxin TANG. Adaptive graph attention Transformer for dynamic traffic flow prediction[J]. Journal of ZheJiang University (Engineering Science), 2025, 59(12): 2585-2592.

[5]	Jinye LI,Yongqiang LI. Spatial-temporal multi-graph convolution for traffic flow prediction by integrating knowledge graphs[J]. Journal of ZheJiang University (Engineering Science), 2024, 58(7): 1366-1376.

[6]	Dian-hai WANG,Rui XIE,Zheng-yi CAI. Prediction of urban interrupted traffic flow based on optimal convergence time interval[J]. Journal of ZheJiang University (Engineering Science), 2023, 57(8): 1607-1617.

[7]	Jing HUANG,Shu-yuan ZHONG,Yuan-qiao WEN,Kun LUO. Adaptive graph generation jump network for traffic flow prediction[J]. Journal of ZheJiang University (Engineering Science), 2021, 55(10): 1825-1833.

[8]	GONG Yue, LUO Xiao-Qin, WANG Dian-hai, YANG Shao-hui. Urban travel time prediction based on gradient boosting regression tress[J]. Journal of ZheJiang University (Engineering Science), 2018, 52(3): 453-460.

Viewed

Full text

Abstract

Cited

Shared

Discussed