Spatio-temporal Travel Volume Prediction and Spatial Dependencies Discovery Using GRU, GCN and Bayesian Probabilities

Travel volumes prediction is an essential topic for urban studies. With accurate estimation, researchers could re-duce the traffic congestion, promote residents’ living quality, and ultimately improve the urban design. Most previous studies only utilised historical travel volume records for prediction, while few of them noticed the impact of spatial factors on travel volumes distributions. In this...