AD-STGCRN: A Model for Traffic Flow Prediction

Abstract Aiming at the problem that existing prediction models fail to utilize serial correlation, temporal correlation and spatial of data from multiple perspectives achieve accurate prediction, a dual-mode spatiotemporal graph convolutional neural network model combined with attention mechanism is proposed. This paper has been tested validated in very stochastic traffic forecasting. experiments on two publicly available datasets, PeMSD04 PeMSD08, show AD-STGCRN higher accuracy, model’s errors reduced by an average 18%, 16% 25% for MAE, RMSE MAPE respectively compared other models. On PeMSD08 they were 20% respectively.