Forecasting of Tourism Demand for Guilin Based on Combined Model

In this paper, we consider the tourism forecasted method to forecast the demand for Guilin. At first, cubic polynomial and GM (1, 1) models are used to forecast tourism demand for Guilin from 1997 to 2010, respectively. Secondly, by comparing the accuracy of cubic polynomial and GM (1, 1) model, a combined model contained a parameter a, a ∈ [0, 1] is proposed. In order to obtain the best parameter value, we construct an optimization problem without restrictive condition and use the Nelder-Mead simplex method to solve it. Finally, mean absolute percentage errors are adopted as criteria for evaluating the accuracy of forecasting exercises. When the parameter a=0.9325, it is shown that the accuracy of presented combined models is higher than cubic polynomial and GM (1, 1) model. Therefore, this kind of combined models is very efficient.