Abstract
Recently, deep learning based city flow prediction has been extensively used in the establishment of smart cities. These methods are data-hungry, making them unscalable to areas lacking data. Although transfer learning can use data-rich source domains to assist target domain cities in city flow prediction, the performance of existing methods cannot meet the needs of actual use, because the long-distance road network connectivity is ignored. To solve this problem, we propose a transfer learning method based on spatiotemporal graph convolution, in which we construct a co-occurrence space between the source and target domains, and then align the mapping of the source and target domains’ data in this space, to achieve the transfer learning of the source city flow prediction model on the target domain. Specifically, a dynamic spatiotemporal graph convolution module along with a temporal encoder is devised to simultaneously capture the concurrent spatiotemporal features, which implies the inherent relationship among the road network structures, human travel habits, and city bike flow. Then, these concurrent features are leveraged as cross-city invariant representations and nonlinearly spanned to a co-occurrence space. The target domain features are thereby aligned with the source domain features in the co-occurrence space by using a Mahalanobis distance loss, to achieve cross-city bike flow prediction. The proposed method is evaluated on the public bike flow datasets in Chicago, New York, and Washington in 2015, and significantly outperforms state-of-the-art techniques.
摘要
最近,基于深度学习的城市流量预测被广泛应用于智慧城市的建设。由于这些方法通常对数据量要求较高,因此难以扩展到数据匮乏的城市。虽然迁移学习可以利用数据丰富的源城市协助目标城市进行城市流量预测,但由于忽略了长距离路网的连通性,因此现有方法的性能无法满足实际使用的需要。为解决这个问题,提出一种基于时空图卷积的迁移预测方法,即在源城市和目标城市之间构建一个共现空间,然后在共现空间中对源城市和目标城市流量数据进行映射对齐,从而实现源城市流量预测模型在目标城市上的迁移预测。具体来说,我们设计了一个动态时空图卷积模块和一个时间编码器,以同时捕捉流量的时间特征和空间特征,这些特征揭示了道路网络结构、人类出行习惯和城市流量之间的内在关联。然后,将这些特征作为跨城市不变表示被非线性映射到共现空间。通过优化马氏距离损失,目标城市特征与源城市特征在共现空间中对齐,从而实现跨城市自行车流量预测。在2015年芝加哥、纽约和华盛顿的公共自行车流量数据集上对所提出的方法进行评估,结果表明该方法明显优于目前最先进的技术。
Similar content being viewed by others
Data availability
The data that support the findings of this study are available from the corresponding authors upon reasonable request.
References
Ahmed MS, Cook AR, 1979. Analysis of freeway traffic time-series data by using Box-Jenkins techniques. Transp Res Rec, 773(722):1–9.
De Maesschalck R, Jouan-Rimbaud D, Massart DL, 2000. The Mahalanobis distance. Chemom Intell Lab Syst, 50(1):1–18. https://doi.org/10.1016/S0169-7439(99)00047-7
Guo SN, Lin YF, Feng N, et al., 2019. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting. 33rd AAAI Conf on Artificial Intelligence, p.922–929. https://doi.org/10.1609/aaai.v33i01.3301922
He ZW, Li Y, Zhang Y, et al., 2023. Ensemble-transfer-learning-based channel parameter prediction in asymmetric massive MIMO systems. Front Inform Technol Electron Eng, 24(2):275–288. https://doi.org/10.1631/FITEE.2200169
Hu J, Lin XH, Wang C, 2022. DSTGCN: dynamic spatial-temporal graph convolutional network for traffic prediction. IEEE Sens J, 22(13):13116–13124. https://doi.org/10.1109/JSEN.2022.3176016
Huang DR, Deng ZP, Zhao L, et al., 2017. A short-term traffic flow forecasting method based on Markov chain and grey Verhulst model. 6th Data Driven Control and Learning Systems, p.606–610. https://doi.org/10.1109/DDCLS.2017.8068141
Huang YJ, Song XZ, Zhang SY, et al., 2021. Transfer learning in traffic prediction with graph neural networks. IEEE Int Intelligent Transportation Systems Conf, p.3732–3737. https://doi.org/10.1109/ITSC48978.2021.9564890
Kang GK, Gao JZ, Chiao S, et al., 2018. Air quality prediction: big data and machine learning approaches. Int J Environ Sci Dev, 9(1):8–16. https://doi.org/10.18178/ijesd.2018.9.1.1066
Li JY, Guo FC, Wang YB, et al., 2020. Short-term traffic prediction with deep neural networks and adaptive transfer learning. IEEE 23rd Int Conf on Intelligent Transportation Systems, p.1–6. https://doi.org/10.1109/ITSC45102.2020.9294409
Lippi M, Bertini M, Frasconi P, 2010. Collective traffic forecasting. European Conf on Machine Learning and Knowledge Discovery in Databases, p.259–273. https://doi.org/10.1007/978-3-642-15883-4_17
Liu J, Guan W, 2004. A summary of traffic flow forecasting methods. J Highway Transp Res Devel, 21(3):82–85.
Lv YS, Duan YJ, Kang WW, et al., 2015. Traffic flow prediction with big data: a deep learning approach. IEEE Trans Intell Transp Syst, 16(2):865–873. https://doi.org/10.1109/TITS.2014.2345663
Ma XL, Dai Z, He ZB, et al., 2017. Learning traffic as images: a deep convolutional neural network for large-scale transportation network speed prediction. Sensors, 17(4):818. https://doi.org/10.3390/s17040818
Mallick T, Balaprakash P, Rask E, et al., 2019. Graph-partitioning-based diffusion convolutional recurrent neural network for large-scale traffic forecasting. Transp Res Rec, 2674(9):473–488. https://doi.org/10.1177/036119812093001
Mallick T, Balaprakash P, Rask E, et al., 2021. Transfer learning with graph neural networks for short-term highway traffic forecasting. 25th Int Conf on Pattern Recognition, p.10367–10374. https://doi.org/10.1109/ICPR48806.2021.9413270
Miao H, Shen JX, Cao JN, et al., 2023. MBA-STNet: Bayes-enhanced discriminative multi-task learning for flow prediction. IEEE Trans Knowl Data Eng, 35(7):7164–7177. https://doi.org/10.1109/TKDE.2022.3179781
Peng H, Wang HF, Du BW, et al., 2020. Spatial temporal incidence dynamic graph neural networks for traffic flow forecasting. Inform Sci, 521:277–290. https://doi.org/10.1016/j.ins.2020.01.043
Seng DW, Lv FS, Liang ZY, et al., 2021. Forecasting traffic flows in irregular regions with multi-graph convolutional network and gated recurrent unit. Front Inform Technol Electron Eng, 22(9):1179–1193. https://doi.org/10.1631/FITEE.2000243
Shao W, Jin ZL, Wang S, et al., 2022. Long-term spatiotemporal forecasting via dynamic multiple-graph attention. 31st Int Joint Conf on Artificial Intelligence, p.2225–2232. https://doi.org/10.24963/ijcai.2022/309
Shao ZZ, Zhang Z, Wang F, et al., 2022. Spatial-temporal identity: a simple yet effective baseline for multivariate time series forecasting. Proc 31st ACM Int Conf on Information & Knowledge Management, p.4454–4458. https://doi.org/10.1145/3511808.3557702
Smola AJ, Schölkopf B, 2004. A tutorial on support vector regression. Stat Comput, 14(3):199–222. https://doi.org/10.1023/B:STCO.0000035301.49549.88
Tascikaraoglu A, 2018. Evaluation of spatio-temporal forecasting methods in various smart city applications. Renew Sustain Energy Rev, 82:424–435. https://doi.org/10.1016/j.rser.2017.09.078
Van Der Voort M, Dougherty M, Watson S, 1996. Combining Kohonen maps with ARIMA time series models to forecast traffic flow. Transp Res Part C Emerg Technol, 4(5):307–318. https://doi.org/10.1016/S0968-090X(97)82903-8
Wang B, Yan Z, Lu J, et al., 2018. Road traffic flow prediction using deep transfer learning. Proc 13th Int FLINS Conf, p.331–338. https://doi.org/10.1142/9789813273238_0044
Wang BW, Wang JS, 2022. ST-MGAT: spatio-temporal multi-head graph attention network for traffic prediction. Phys A, 603:127762. https://doi.org/10.1016/j.physa.2022.127762
Wang LY, Geng X, Ma XJ, et al., 2019. Cross-city transfer learning for deep spatio-temporal prediction. Proc 28th Int Joint Conf on Artificial Intelligence, p.1893–1899. https://doi.org/10.24963/ijcai.2019/262
Wang SZ, Miao H, Li JY, et al., 2022. Spatio-temporal knowledge transfer for urban crowd flow prediction via deep attentive adaptation networks. IEEE Trans Intell Transp Syst, 23(5):4695–4705. https://doi.org/10.1109/TITS.2021.3055207
Williams BM, Hoel LA, 2003. Modeling and forecasting vehicular traffic flow as a seasonal ARIMA process: theoretical basis and empirical results. J Transp Eng, 129(6):664–672.
Xiang SM, Nie FP, Zhang CS, 2008. Learning a Mahalanobis distance metric for data clustering and classification. Patt Recog, 41(12):3600–3612. https://doi.org/10.1016/j.patcog.2008.05.018
Xu ZY, Kang Y, Cao Y, 2023. High-resolution urban flows forecasting with coarse-grained spatiotemporal data. IEEE Trans Artif Intell, 4(2):315–327. https://doi.org/10.1109/TAI.2022.3153750
Yao ZX, Xia SC, Li Y, et al., 2023. Transfer learning with spatial-temporal graph convolutional network for traffic prediction. IEEE Trans Intell Transp Syst, 24(8):8592–8605. https://doi.org/10.1109/TITS.2023.3250424
Yu B, Yin HT, Zhu ZX, 2018. Spatio-temporal graph convolutional networks: a deep learning framework for traffic forecasting. Proc 27th Int Joint Conf on Artificial Intelligence, p.3634–3640. https://doi.org/10.24963/ijcai.2018/505
Zhao L, Song YJ, Zhang C, et al., 2020. T-GCN: a temporal graph convolutional network for traffic prediction. IEEE Trans Intell Transp Syst, 21(9):3848–3858. https://doi.org/10.1109/TITS.2019.2935152
Author information
Authors and Affiliations
Contributions
Binkun LIU and Zhenyi XU designed the research. Binkun LIU processed the data and drafted the paper. Yang CAO helped organize the paper. Yu KANG and Yunbo ZHAO helped in data control and project management. Yang CAO and Zhenyi XU revised and finalized the paper. Yu KANG and Zhenyi XU provided the funding acquisition.
Corresponding authors
Ethics declarations
All the authors declare that they have no conflict of interest.
Additional information
Project supported by the National Natural Science Foundation of China (Nos. 62103124 and 62033012), the Major Special Science and Technology Project of Anhui Province, China (No. 202003a07020009), and the Open Project Program of Key Laboratory of Ministry of Education of System Control and Information Processing, China (No. SCIP20230109)
Rights and permissions
About this article
Cite this article
Liu, B., Kang, Y., Cao, Y. et al. Transfer learning with a spatiotemporal graph convolution network for city flow prediction. Front Inform Technol Electron Eng 26, 79–92 (2025). https://doi.org/10.1631/FITEE.2300571
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1631/FITEE.2300571