Smart City Gnosys

Smart city article details

Title	Traffic Flow Prediction Based On Adjacency Graph And Attention Mechanism
ID_Doc	58579
Authors	Zhao J.
Year	2024
Published	2024 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024
DOI	http://dx.doi.org/10.1109/AINIT61980.2024.10581791
Abstract	Traffic flow prediction plays a crucial role in smart city systems and serves as a fundamental component for various traffic applications. The primary challenge lies in effectively modeling the spatiotemporal dependencies of traffic flow. Current methods mainly employ Graph Convolutional Networks (GCN) to model spatial relationships and adopt Convolutional Neural Networks (CNN) or Recurrent Neural Networks (RNN) to capture temporal dependencies. However, existing spatial modeling methods often only utilize adjacency matrices to represent local relationships, neglecting the global spatial information. Within the road network, certain roads exhibit similar surrounding network structures and perform similar roles, making their features valuable for traffic flow prediction. In this paper, we propose MA-STGCN, a spatiotemporal network model based on multiple adjacency graphs and a multi-head attention mechanism. The model includes: 1) utilizing the node2vec algorithm to obtain vector representations of roads in the road network, and calculating a similarity matrix through a threshold to enable graph convolutional operations for extracting global spatial information; 2) employing a multi-channel self-attention mechanism to delve into the spatiotemporal features of the model. Experimental results conducted on the production datasets KM - Data and CQ- Data verify the effectiveness of the proposed model, demonstrating significant improvements in accuracy compared to mainstream models. © 2024 IEEE.
Author Keywords	Attention mechanism; Dilated convolution; node embedding; Traffic forecasting

Similar Articles

Id	Similarity	Authors	Title	Published
53044	0.947	Meng X.; Xie W.; Cui J.	Stmgfn: Spatio-Temporal Multi-Graph Fusion Network For Traffic Flow Prediction	Lecture Notes in Computer Science, 15291 LNCS (2025)
11043	0.945	He Q.; Xia D.; Li J.; Yang J.; Hu Y.; Li Y.; Li H.	Attention-Based Spatiotemporal Adaptive Graph Diffusion Convolutional Network For Traffic Flow Prediction	Transportation Research Record (2025)
52517	0.945	Huang X.; Pan Z.; Zhao G.	Spatial-Temporal Interactive Graph Convolutional Networks For Traffic Forecasting	2024 4th International Conference on Electronic Information Engineering and Computer Technology, EIECT 2024 (2024)
11029	0.944	Zhang H.; Liu J.; Tang Y.; Xiong G.	Attention Based Graph Covolution Networks For Intelligent Traffic Flow Analysis	IEEE International Conference on Automation Science and Engineering, 2020-August (2020)
58588	0.939	Wang X.	Traffic Flow Prediction Method Based On Deep Learning And Graph Neural Network	International Journal of High Speed Electronics and Systems (2025)
4883	0.938	Cao S.; Wu L.; Wu J.; Wu D.; Li Q.	A Spatio-Temporal Sequence-To-Sequence Network For Traffic Flow Prediction	Information Sciences, 610 (2022)
52568	0.938	Ren H.; Kang J.; Zhang K.	Spatio-Temporal Graph-Tcn Neural Network For Traffic Flow Prediction	2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing, ICCWAMTIP 2022 (2022)
38024	0.938	Yao H.; Chen R.; Xie Z.; Yang J.; Hu M.; Guo J.	Mra-Dgcn: Multi-Range Attention-Based Dynamic Graph Convolutional Network For Traffic Prediction	Proceedings - 2022 IEEE International Conference on Big Data, Big Data 2022 (2022)
58576	0.935	Zheng L.; Pu Y.; Sun W.	Traffic Flow Prediction Algorithm Based On Attention Spatiotemporal Graph Convolution Mechanism	IEEE Advanced Information Technology, Electronic and Automation Control Conference (IAEAC) (2024)
38047	0.934	Yang S.; Wu Q.; Wang Y.; Zhou Z.	Mstdfgrn: A Multi-View Spatio-Temporal Dynamic Fusion Graph Recurrent Network For Traffic Flow Prediction	Computers and Electrical Engineering, 123 (2025)