Study on the Forecast of Air Passenger Flow Based on SVM Regression Algorithm

The forecast of air passenger flow plays an important role in the management of airline, but the traditional forecast methods can´t guarantee the generalization capability when they face a large-scale, multi-dimension, nonlinear and non-normal distribution time series data. To improve the forecast ability of air passenger flow, the SVM regression algorithm is introduced in this paper. By selecting appropriate parameters and kernel function, compared with the other two forecast methods, we find that the result obtained by SVM regression algorithm shows the least error among the mentioned three methods.