Advances in spatiotemporal graph neural network prediction research

Being a kind of non-Euclidean data, spatiotemporal graph data exists everywhere from traffic flow, air quality index to crime case, etc. Unlike the raster irregular and disordered characteristics have attracted research interest scholars, with prediction being one hot spots. The emergence neural networks (ST-GNNs) provides new insight for solving problem obtaining spatial correlation while achieving state-of-the-art performance. In this paper, comprehensive survey on ST-GNNs domain is presented, where background introduced before computational paradigm ST-GNN thoroughly reviewed. From perspective model construction, 59 well-known models in recent years are classified discussed. Some these further analyzed terms performance efficiency. Subsequently, categories application fields summarized, providing clear idea technology selection different applications. Finally, evolution history future direction also facilitate researchers timely understand current state by ST-GNNs.