兵工自动化2025，Vol.44Issue(5)：66-70,5.DOI:10.7690/bgzdh.2025.05.015
动静图融合和时序流注意力网络用于交通流预测
Dynamic-static Graph Fusion and Temporal Flow Attention Network for Traffic Flow Prediction
摘要
Abstract
Accurately predicting traffic flow is conducive to optimizing traffic management and improving traffic efficiency,a new dynamic-static graph fusion and temporal flow attention network is proposed.The graph convolutional network is used to capture dynamic and static spatial correlations.A flow attention mechanism is introduced to effectively alleviate the quadratic complexity problem.A temporal correlation modeling(TCM)module is designed to replace the linear transformation method of the flow attention mechanism,so as to enhance the model's temporal modeling ability.A large number of experiments are carried out on four real-world traffic datasets.The results show that the proposed model has superior performance and significantly outperforms the baselines.关键词
交通流预测/时空相关性/流注意力机制/图卷积网络/特征融合Key words
traffic flow prediction/spatio-temporal correlation/flow attention mechanism/graph convolutional network/feature fusion分类
信息技术与安全科学引用本文复制引用
闫敬,王祥,郑铮..动静图融合和时序流注意力网络用于交通流预测[J].兵工自动化,2025,44(5):66-70,5.基金项目
唐山师范学院科研基金项目(2022C53) （2022C53）
