Spatio-temporal graph attention networks for traffic prediction

Abstract

ABSTRACTThe constraints of road network topology and dynamically changing traffic states over time make the task of traffic flow prediction extremely challenging. Most existing methods use CNNs or GCNs to capture spatial correlation. However, convolution operator-based methods are far from optimal in their ability to fuse node features and topology to adequately model spatial correlation. In order to model the spatio-temporal features of traffic flow more effectively, this paper proposes a traffic flow prediction model, the Spatio-Temporal Graph Attention Network (STGAN), which is based on graph attention mechanisms and residually connected gated recurrent units. Specifically, a graph attention mechanism and a random wandering mechanism are used to extract spatial features of the traffic network, and gated recurrent units with residual connections are used to extract temporal features. Experimental results on real-world public transportation datasets show that our approach not only yields state-of-the-art performance, but also exhibits competitive computational efficiency and improves the accuracy of traffic flow prediction.KEYWORDS: Traffic flow predictiongraph attention mechanismresidual connectionneural networks AcknowledgmentsThis work is supported by the National Natural Science Foundation of China (Grant No. 62272120, 62106030); the Technology Innovation and Application Development Projects of Chongqing (Grant No. cstc2021jscx-gksbX0032, cstc2021jscx-gksbX0029); the Research Program of Basic Research and Frontier Technology of Chongqing (Grant No. cstc2021jcyj-msxmX0530); the Key R\& D plan of Hainan Province (Grant No. ZDYF2021GXJS006).Disclosure statementNo potential conflict of interest was reported by the author(s).Additional informationFundingThe work was supported by the National Natural Science Foundation of China [62272120, 62106030]; Research Program of Basic Research and Frontier Technology of Chongqing [cstc2021jcyj-msxmX0530]; Key R & D plan of Hainan Province [ZDYF2021GXJS006]; Technology Innovation and Application Development Projects of Chongqing [cstc2021jscx-gksbX0032, cstc2021jscx-gksbX0029].