Research On The Cause Determination And Prediction Of Flight Delay

The harm caused by flight delay is becoming more and more serious.How to alleviate delay has become an urgent problem to be solved.It is necessary to start from the source,study the flight normality regulations,explore the causes of delay,and predict the flight delay.However,there are differences in the specific implementation of normality judgment among various units.There is less research on the identification of delay causes and the clarification of responsibilities.In many cases,complete flight information cannot be collected,which affects the accuracy of delay prediction.In this context,based on the data collection and fusion processing,the comparative analysis of normality statistical methods,delay cause identification and prediction are carried out.Firstly,the flight operation data and meteorological data were collected,and they were fused according to the key time nodes and information to restore the data of the whole flight operation process.Then the transformation and normalization of characteristic indexes were carried out to lay a foundation for subsequent research.Secondly,the contents of different versions of normality statistics methods were studied,combined with European and American for comparative analysis.And the flight delay time in different months was calculated according to the fusion data and the formulas in different versions,used to analyze the gap between different versions and find the existing laws and put forward reasonable suggestions.Thirdly,the causes of flight delay were studied,the frequency of delay causes were counted in combination with the actual data.The XGBoost model was established,and the time spent between multiple key links was added as the characteristic index.After solving the imbalance of cause categories,the causes of flight delay were determined and compared with other algorithms.The results show that the algorithm has high accuracy.Finally,the prediction of flight delay was studied.The GA-Light GBM model was constructed,adding the previous delay information,delay reasons and associated flow control information as characteristic indicators to predict the departure delay.Compared with other algorithms,the results show that the algorithm has better prediction accuracy.Figure 19;Table 19;Reference 48...