Support vector machine-based combinational model for air traffic forecasts
Bing Xu · Journal of Tsinghua University(Science and Technology) · 2008
A flow forecasting auto-regression model based on the support vector machine(SVM) was developed to improve the accuracy of air traffic forecasts.The SVM model was then combined with a polynomial auto-regression model and a robust auto-regression model into a combined forecasting model.The SVM,polynomial and robust auto-regression forecasting models and the combined forecasting model were tested with real air traffic data from the Beijing ATC Region.The results show that the SVM forecasting errors are less than 5% and the combined model's errors are less than 2%,thus the combined forecasting method gives the best results among these methods.