GLGAT: Global-Local Graph Attention Network For Traffic Forecasting

Spatiotemporal prediction plays an important role in many area, especially intelligent transportation system (ITS). However, due to the complex spatiotemporal conditions, traffic flow prediction task is challenging. In this paper, we propose a novel GLGAT model to predict spatiotemporal traffic flow. GLGAT can grasp global and local spatiotemporal information respectively, and make accurate predictions of the traffic flow. We test GLGAT on PeMSD7 data set and the experiment results indicate that GLGAT could outperform the state-of-the-art baselines especially in long-term prediction.