Traffic Prediction Using Graph Convolution Networks

Authors

  • Bathina Bhanukowshik   Department of Computer Science and Engineering, National Institute of Technology, Andhra Pradesh, India

Keywords:

Traffic prediction, Graph convolutional network (GCN), Deep learning, Spatial-temporal data, Time series forecasting, Hyperparameter optimization, Model evaluation, Accuracy, Precision, Recall, F1 score, Root mean squared error (RMSE), Mean absolute error (MAE), Mean absolute percentage error (MAPE).

Abstract

Traffic Prediction has become very essential now a days. It is helpful to manage the traffic in some major cities. In reality, the traffic data generated contains some missing values, due to communication errors, sensor faults, etc. Since the traffic data is Spatial-Temporal, which is very complex, it becomes difficult to handle these situations. So, the authors propose a Graph Convolution Network-based model to predict future traffic with missing values. The missing value processing and traffic prediction are processed in one step. The model considers both Spatial-Temporal data and the history of the particular junction traffic. The computational time complexity of the network is optimized for predicting the traffic.

References

  1. Graph convolutional networks for traffic forecasting with missing values, Jingwei Zuo, Karine Zeitouni, Yehia Taher & Sandra Garcia-Rodriguez, published in DMKD 2022.
  2. Informer: Beyond efficient transformer for long sequence time-series forecasting, Zhou H, Zhang S, Peng J, et al, published in AAAI 2021.
  3. Diffusion convolutional recurrent neural network: Data-driven traffic forecasting, by Li Y, Yu R, Shahabi C, et al published in International Conference on Learning Representations (ICLR), 2018.
  4. Graph wavenet for deep spatial-temporal graph modeling by Wu Z, Pan S, Long G, et al published in Proceedings of the 28th International Joint Conference on Artificial Intelligence (IJCAI), pp 1907 – 1913, 2018.
  5. Dynamic graph convolutional recurrent network for traffic prediction: Benchmark and solution by Li F, Feng J, Yan H, et al published in ACM Transactions on Knowledge Discovery from Data (TKDD) , 2021.
  6. Che Z, Purushotham S, Cho K, et al Recurrent neural networks for multivariate time series with missing values published in Scientific reports 8 – 12 , 2018.
  7. Dataset : METR-LA, downloaded from https://drive.google.com/drive/folders/10FOTa6HXPqX8Pf5WRoRwcFnW9BrNZEIX
  8. Node2Vec: Scalable Feature Learning for Networks by Aditya Grover, Jure Leskovec published in proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016.
  9. Semi-Supervised Classification with Graph Convolutional Networks by Thomas N. Kipf, Max Welling, published in ICLR, 2017.
  10. Deep Learning on Graphs: A Survey, by Zhang, Z., Cui, P., Zhu, W., & Wang, X. published in IEEE Transactions on Knowledge and Data Engineering, 30(12), 2321-2333., 2018.
  11. Graph Convolutional Long Short-Term Memory Network for Traffic Flow Forecasting by Lv, Z., Zhu, Y., & Wu, L., published in IEEE Transactions on Intelligent Transportation Systems, 21(3), 1074-1084., 2020.
  12. Traffic Prediction with Graph Convolutional Recurrent Neural Network by Ma, X., Dai, Z., He, Z., Ma, J., Wang, Y., & Wang, Y., published in the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 615-623), 2019.

Downloads

Published

2023-08-30

Issue

Section

Research Articles

How to Cite

[1]
Bathina Bhanukowshik , " Traffic Prediction Using Graph Convolution Networks" International Journal of Scientific Research in Computer Science, Engineering and Information Technology(IJSRCSEIT), ISSN : 2456-3307, Volume 9, Issue 4, pp.57-65, July-August-2023.