Abstract
Traffic prediction could successfully reduce the environmental pollution and exhaust emissions brought on by traffic congestion while offering trustworthy travel advice for people's everyday commutes. However, because of the intricate dynamic Spatio-temporal linkage of traffic data. Therefore, it is difficult to anticipate traffic flow accurately. Past research has frequently combined Spatio-temporal prediction models with Graph Neural Network (GNN) and temporal processing modules to capture traffic networks' spatial and temporal dependence. However, they still frequently experience the following drawbacks: (1) relying solely on static spatial connectivity relationships between traffic road nodes, the dynamic impact of traffic timing data on spatial dependence is ignored; (2) losing the global temporal dependence information quickly when extracting long-term traffic sequence data using recursively structured temporal components like Recurrent Neural Network (RNN). In order to achieve this, we proposed the Dynamic Spatio-temporal Graph Recurrent Network (DSTGRN), which accomplishes fine-grained modeling of the temporal dependence of traffic flow data by encoding road nodes and using both spatial attention mechanism and multi-headed temporal attention mechanism. Experimental results on traffic datasets from the real world indicate that our proposed model outperforms its baseline model in terms of accuracy of prediction.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.