Flight Delay Prediction Based On Long Short-Term Memory And Residual Network

With the development of the civil aviation industry,flight delay has gradually become a hot topic.Flight delay has caused considerable worry to our daily trips.Therefore,more accurate flight delay prediction can help people to reasonably arrange travel plans.Neural network has been used in the study of flight delay prediction,but the existing neural network model is simple in structure,and the influence factors are not comprehensive enough,so the feature expression ability is limited,which makes the accuracy of flight delay prediction not high enough.In view of the shortcomings of the existing neural network model,we propose a flight delay prediction model based on long-term and short-term memory and residual network(LSTM-ResNet)in this paper.The model includes a long-term and short-term memory network(LSTM)structure in a recurrent neural network(RNN)and a residual network(ResNet)structure in a convolutional neural network(CNN).The LSTM structure in this model can extract the hidden time series features in flight information.The ResNet structure can retain the features in the shallow network on the basis of deepening the network layers,so that the features will not be lost in the deep network,thus solving the problem of gradient disappearance caused by deepening the network structure.Compared with the existing neural network model,more influence factors of the model is considered,such as the influence of weather conditions on flight delays,and the data fusion between flight information and weather information is processed to further enhance ability of the network model in extracting data features.The flight delay prediction model proposed in this paper is compared with the traditional method for single flight data set and single airport data set respectively.The experimental results show that the prediction accuracy of the model is improved compared with the traditional method.Finally,a flight delay prediction platform is implemented in this paper,which can predict the delay level of future flights.