我国入境旅游人次月度指数预测模型比较研究

A Comparison Study on the Index Forecast Model of Monthly Inbound Tourists in China

需求预测是旅游产业经营决策的基本依据,但产业的广泛关联性与各类突发事件使旅游需求预测尤其是中短期预测较为困难。本文采用X12-ARIMA模型、TRAMO/SEATS模型、ARMA模型与GARCH模型,对异常数据点采用附加的外部冲击调整,利用7种估计方法估计了我国入境旅游人次的月度指数并进行了预测比较,发现采用外部冲击检测的TRAMO/SEATS模型由于能有效提取序列数据的信息,对预测我国入境旅游人次最为有效。

Demand forecast is the fundamental basis of the decisionmaking for the operation of tourism industries, but the wide interconnection to other industries and all kinds of imminent events make it more difficult to forecast tourism demand, especially for mid-and short-term prediction. Based on X12-ARIMA model, TRAMO/SEATS model, ARMA model and GARCH model, with additive external adjustment for extraordinary data, the paper makes forecasting comparison of monthly index of inbound tourists to China by using seven different forecasting methods. It is found that adopting TRAMO/SEATS model with AO automatic detection is the most effective way for inbound tourist forecast because of the efficient collection of sequence data.