Research Article
Geographic and Temporal Deep Learning Method for Traffic Flow Prediction in Highway Network
@INPROCEEDINGS{10.1007/978-3-030-92638-0_23, author={Tianpu Zhang and Weilong Ding and Mengda Xing and Jun Chen and Yongkang Du and Ying Liang}, title={Geographic and Temporal Deep Learning Method for Traffic Flow Prediction in Highway Network}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 17th EAI International Conference, CollaborateCom 2021, Virtual Event, October 16-18, 2021, Proceedings, Part II}, proceedings_a={COLLABORATECOM PART 2}, year={2022}, month={1}, keywords={Deep learning Graph convolutional network Long short-term memory Traffic flow prediction Spatio-temporal data}, doi={10.1007/978-3-030-92638-0_23} }- Tianpu Zhang
Weilong Ding
Mengda Xing
Jun Chen
Yongkang Du
Ying Liang
Year: 2022
Geographic and Temporal Deep Learning Method for Traffic Flow Prediction in Highway Network
COLLABORATECOM PART 2
Springer
DOI: 10.1007/978-3-030-92638-0_23
Abstract
Traffic congestion has become an inevitable situation faced by all countries and the prediction accuracy of traffic flow, as one of the means to solve this problem, still needs to be improved. Most studies lack the consideration of the influence of multiple factors such as spatial factors, time series factors and other external factors, which makes the prediction effect of traffic flow unsatisfactory. In this paper a method is proposed based on deep learning that can capture the geographic spatial relationship among toll stations, the dynamic temporal relationship of historical traffic flow, extreme weather and calendar types. On the three metrics of MAPE, MAE, and RMSE, the prediction effect of our model has increased by 30% compared with KNN, GBRT and LSTM models.
