Intelligent Prediction Technology for Air Traffic Flow Based on Multi-Scale Spatiotemporal Information
In order to solve the problem of spatiotemporal complexity in air traffic flow prediction,an intelligent prediction method for air traffic flow based on multi-scale spatiotemporal information is pro-posed.The irregular spatial relationships in air traffic networks and the periodic and trending character-istics of ultra-long time series are modeled.The multi-dimensional criticality of air traffic flow is cap-tured by constructing the graph convolution layers and the stacked dilated causal convolutional layers,and then the accurate prediction of future traffic flow is provided.The experimental results show that the prediction accuracy of the method is higher than the traditional method on multiple aviation net-work data sets,and it has practical application value.Thus,it can provide a reference for intelligent decision-making of air traffic management system.
air traffic flow predictionmulti-scale spatiotemporal informationgraph representation learninggraph convolutional network
NETL
NSTL
万方数据
维普
