A High-precision Method of Flight Arrival Time Estimation based on XGBoost

Aircraft's estimated time of arrival is an important issue in flight operation, many methods were used to get a higher accuracy. Deep learning methods can offer great favor to improve it. In this paper, a high-precision method of flight arrival time estimation based on XGBoost regression. First, Historical data should be processed by using correlation coefficient analysis of the data features, and features that are highly correlated with the flight arrival time are determined; then based on a large number of historical flight operation data, the selected features of historical data are input into the training based on XGBoost regression to build a flight arrival time prediction model. And finally input the target flight's real-time information of the selected features to get the remaining flight time of the flight. At the end of this paper, an experiment was given to analyze the accuracy of the prediction model. With the analysis of experiment's results, the new method this paper proposed shows a good performance.