A Novel Attention Ensemble Traffic Flow Prediction Model based on Spatio-Temporal big data

The emergence of intelligent transportation systems has put forward higher requirements for the problem of traffic flow prediction. Since the topology of the traffic road network changes dynamically, the traditional convolutional neural network that extracts the features of graph structure data cannot capture such dynamic features. In this paper we propose a traffic flow prediction model based on spatiotemporal characteristics to predict the traffic speed in the traffic road network. The dynamic topology structure in the traffic road network is obtained through the graph attention network (GAT), and the temporal characteristics of the data are processed by the long short-term memory network (LSTM), and the periodic characteristics of the traffic flow are captured. In this paper, the prediction model is mainly divided into space module and time module, which can capture the characteristics of traffic data in an all-round way, reduce the loss function value, and improve the prediction accuracy of the model.