Dynamic Prediction about Turnaround Time of Flight based on Support Vector Machine Regression

Validly predicting turnaround time of flight will assist management department of the airport to take corresponding measures and to improve the operational efficiency of the airport. In this paper, the method of Support Vector Machine Regression (SVR) is utilized to establish a prediction model about turnaround time of flight. Firstly, the time of arrival delay about the former flight is called as internal factors, and combine with the time of actual time about the current flight to construct input variable of the model. Secondly, utilizing the Gaussian Radial Basis Function to select the optimal parameters. Finally, utilizing existing historical data to construct training model, and compare the prediction performance of single factor as input variable with multiple factors as input variable respectively. The results demonstrate that the model of this paper derived has a minimum error and an accurate estimation about turnaround time.