Study on Subway passenger flow prediction based on deep recurrent neural network


As the construction and management of subway transit system becomes increasingly mature, analyzing the passenger flow information of the normal transportation network and accurately predicting the passenger flow in a short time have become the core of subway transit system operation and management. However, it is difficult for traditional intelligent prediction algorithms to meet the high accuracy and fast response capabilities required for predicting passenger flow in a short time in unexpected situations. In order to improve the prediction performance, this paper proposes a time series prediction model based on deep recurrent neural network (DRNN). Using DRNN’s unique memory function to capture the dynamic information of the time series, we can better learn the “trend” between data at different moments, so that we can more accurately predict the output at the next moment. The comparison among the case studies based on the measured data of subway passenger flow with time series characteristics, the traditional support vector machine and the neural network method, shows that DRNN prediction has the smallest overall deviation, small deviation fluctuation and good robustness.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6


  1. 1.

    Arem BV, Kirby HR, Vlist MJMVD et al (1997) Recent advances and applications in the field of short-term traffic forecasting. Int J Forecast 13(1):1–12

    Article  Google Scholar 

  2. 2.

    F Beaufays, H Sak, A Senior (2014) Long Short-Term Memory Recurrent Neural Network Architectures for Large Scale Acoustic Modeling. Interspeech, no. September 338–342

  3. 3.

    Chai Y, Huang D, Zhao L (2016) A short-term traffic flow prediction method based on wavelet analysis and neural network. Control and decision conference. IEEE 7030-1034.

  4. 4.

    Dia H (2001) An object-oriented neural network approach to short-term traffic forecasting. Eur J Oper Res 131(2):253–261

    Article  Google Scholar 

  5. 5.

    Dong Shengwei (2013) Study on short-term passenger flow prediction of rail transit based on improved BP neural network. Beijing Jiaotong University

  6. 6.

    Du Y, Wang W, Wang L. (2015) Hierarchical recurrent neural network for skeleton based action recognition [J]. Journal of intelligent perception and computing research center (CVPR):1110-1118

  7. 7.

    Enjian Y, Wenhua Z, Yongsheng Z (2018) Prediction of real-time inbound and outbound passenger flow of the new stations on urban rail transit lines in the initial stage of opening. Chin Railw Sci 02:119–127

    Google Scholar 

  8. 8.

    Gensuo M, Liang L, Yang R (2015) Application of Grey variation particle swarm algorithm in prediction of bus passenger flow. Comput Eng Sci 37(1):104–110

    Google Scholar 

  9. 9.

    Gooijer JGD, Hyndman RJ (2005) 25 years of time series forecasting. Monash Econ Bus Stat Work Pap 22(3):443–473

    Google Scholar 

  10. 10.

    A Graves, A Mohamed, G Hinton (2013) Speech recognition with deep recurrent neural networks. 2013 IEEE Int. Conf. Acoust. Speech Signal Process (6): 6645–6649

  11. 11.

    K Gregor, A Graves, WG Com (2015) DRAW : A Recurrent Neural Network For Image Generation. Proceeding of the 32nd International Conference on Machine Learning, JMLR: W&CP 37

  12. 12.

    Guo et al (2019) Short-term abnormal passenger flow prediction based on the fusion of SVR and LSTM. IEEE Access 7:42946–42955

    Article  Google Scholar 

  13. 13.

    Han et al (2019) Short-Term Prediction of Bus Passenger Flow Based on a Hybrid Optimized LSTM Network. ISPRS Int J Geo Inf 8(9):366

    Article  Google Scholar 

  14. 14.

    Jie L, Yuming H, Hui Z, Xuejin G, Pu W (2014) A Combined Model for Predicting Short-term Passenger Flow in Urban Transportation Hubs. Trans Info Safety 32(02):41–44+49

    Google Scholar 

  15. 15.

    U Lowell, J Donahue, UC Berkeley, M Rohrbach, UC Berkeley, and R Mooney (2015) Translating Videos to Natural Language Using Deep Recurrent Neural Networks. Human Language Technologies: The 2015 Annual Conference of the North American Chapter of the ACL 1494–1504

  16. 16.

    Ma S, Wuniriqiqige, Xiaoping L (2016) A review of big data and deep learning. Trans Intel Syst 11(6):728–742

    Google Scholar 

  17. 17.

    R Pascanu, C Gulcehre, K Cho, Y Bengio (2013) How to construct deep recurrent Neur Networks 1–13

  18. 18.

    Smith BL, Williams BM, Oswald RK (2002) Comparison of parametric and nonparametric models for traffic flow forecasting. Transp Res C 10(4):303–321

    Article  Google Scholar 

  19. 19.

    Wang H, Yan L (2016) Predictingof metro passenger flow by combined Grey support vector machines. J Dalian Jiaotong Univ 37(2):6–9

    MathSciNet  Google Scholar 

  20. 20.

    Xiaoxue Y, Zheng Y, Lin J (2016) Passenger flow prediction of urban rail transit based on improved WNN. Comput Eng Appl 52(11):227–232

    Google Scholar 

  21. 21.

    Zhu et al (2019) Prediction of passenger flow in urban rail transit based on big data analysis and deep learning. IEEE Access 7:142272–142279

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Deqiang Liu.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Foundation Project: project supported by the National Natural Science Foundation of China (71771151)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Liu, D., Wu, Z. & Sun, S. Study on Subway passenger flow prediction based on deep recurrent neural network. Multimed Tools Appl (2020).

Download citation


  • Subway passenger flow
  • Passenger flow prediction
  • Deep recurrent neural network
  • Time series prediction