APS-PBW: The Analysis and Prediction System of Customer Flow Data Based on WIFI Probes

  • Yuanyuan Wu
  • Shunhua Gu
  • Tong Yu
  • Xiaolong XuEmail author
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11062)


The collection, analysis and prediction of the customer flow data can provide all-dimensional data reference for the refined operating of the enterprise. In the meantime, analysis system of the customer flow data can not only help the enterprise detect the marketing effectiveness, but also discover potential opportunities and improvement measures, providing all-round data reference for the efficient and sophisticated operation of the enterprise. We build the analysis and prediction system of customer flow data based on WIFI probe (APS-PBW). APS-PBW takes the WIFI probe as the data collector, which can scan the mobile devices within its range during a short time interval, and also get the information about the MAC addresses, the reference distances and the time stamps of mobile phones. Then, we do some statistical analyses for the indexes of the records from the customer’s angle and the store’s angle, which include the length of the customer’s entry, the cycle of the customer’s visit, the customer flow of the store, the number of new and old customers, etc. Meanwhile, SARIMA model and BP neural network model are applied to the system to predict the customer flow data respectively. To conclude, the framework of our system can be divided into three parts: the collection of customer flow data based on the WIFI probe, the analysis and prediction of the customer flow data by the means of SARIMA model and BP neural network model, and the system construction. We implement a series of experiments to test the performance of the prediction system about the customer flow data. The experimental results show that, compared with BP neural network model, SARIMA model is more suitable and also more accurate for the prediction of the customer flow data.


WIFI probe SARIMA model BP neural network model Customer flow data prediction 



We would like to thank the reviewers for their comments, which helped us significantly improve the quality of this paper. And this work is supported by the National Natural Science Foundation of China under Grants 61472192 and 61772286, the National Key Research and Development Program of China under Grant 2018YFB1003702, the Scientific and Technological Support Project (Society) of Jiangsu Province under Grant BE2016776, the “333” project of Jiangsu Province under Grants BRA2017228 and BRA2017401, the Talent Project in Six Fields of Jiangsu Province under Grant 2015-JNHB-012, and the Science and Technology Innovation Training Program.


  1. 1.
    Du, P.: Algorithm improvement of parameter estimation for ARMA model and application of SARIMA model. M.S. thesis, Yanshan University, China (2010)Google Scholar
  2. 2.
    Eni, D., Adeyeye, F.J.: Seasonal ARIMA modeling and forecasting of rainfall in Warri town. Nigeria. J. Geosci. Environ. Protect. 3(6), 91–98 (2015)CrossRefGoogle Scholar
  3. 3.
    Gao, J.: Artificial Neural Network Theory and Simulation Examples. China Machine Press, Beijing (2003)Google Scholar
  4. 4.
    Guo, Y., Song, X.: Analysis and improvement of gradient descent method. Technol. Outlook 26(15), 115–117 (2016)Google Scholar
  5. 5.
    Hang, Z.: Research and application of traffic statistics algorithm in intelligent video monitoring. M.S. thesis, Lanzhou University of Technology, China (2016)Google Scholar
  6. 6.
    Luo, B., Huang, W.J., Yang, S.: Improved algorithm of BP neural networks based on parameters adjustment of tan-sigmoid transfer function. J. Chongqing Univ. 29(1), 150–156 (2006)Google Scholar
  7. 7.
    Mills, T.: The Econometric Modelling of Financial Time Series, vol. 96. Cambridge University Press, Cambridge (1996)zbMATHGoogle Scholar
  8. 8.
    Hecht-Nielsen, R.: Kolmogorov’s mapping neural network existence theorem. In: 1987 IEEE International Conference on Neural Network, pp. 11–13 (1987)Google Scholar
  9. 9.
    Shen, H., Wang, Z., Gao, C., Qin, J., Yao, F., Xu, W.: Determining the number of BP neural network hidden layer units. J. Tianjin Univ. Technol. 24(5), 13–15 (2008)Google Scholar
  10. 10.
    Wang, Y.: Applied Time Series Analysis. China Renmin University Press, Hebei (2015)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Jiangsu Key Laboratory of Big Data Security and Intelligent ProcessingNanjing University of Posts and TelecommunicationsNanjingChina
  2. 2.Institute of Big Data Research at YanchengNanjing University of Posts and TelecommunicationsYanchengChina

Personalised recommendations