Optimal weighted combinatorial forecasting model in the airport cargo prediction

According to the characteristics of the airport cargo, exponential curve model, unitary linear regression model and GM(1,1) model for the S individual airport cargo forecasts are conducted with good prediction, while the average absolute error of each model is no more than 5%. According to the criterion of minimizing the three single forecasting models´ forecasting error sum of squares, we establish an optimal weighted combinatorial forecasting model, and use it to forecast the airport cargo with the average absolute error of only 1.73%, which further improves the predicting accuracy of the airport cargo.