Abstract: ARIMA model has a better fitting effect on seasonal feature. Grey model can accurately reflect the growth trend of time series. Combined the characteristics of civil aviation RFTK with the advantages of ARIMA model and GM（1, 1） model, comprehensively using the analysis method of time series, this paper presents the combination prediction model of ARIMA-GM. It respectively establishes the time series model of ARIMA model and GM（1, 1） model to show the dynamic rule of civil aviation RFTK changing as time passing. At last, for more accurately predicting the month civil aviation RFTK, this paper puts forward the forecasting method of combination model, and makes an exact forecast of the civil aviation RFTK in a few months. The combination model is a good solution to the practical problems for civil aviation RFTK forecasting, based on which it can have a macro-grasp of civil aviation RFTK market trend, which will certainly be conducive to economic decision-making.

Key words: civil aviation RFTK, GM（1, 1） model, Autoregressive Integrated Moving Average（ARIMA） model, combination model

摘要： ARIMA模型对季节特征有较好的拟合效果，灰色GM（1，1）模型能准确反映时间序列的增长趋势，结合民航货邮周转量的特点和ARIMA模型和GM（1，1）模型的优点，分别建立货邮周转量的ARIMA和GM（1，1）的时间序列模型，揭示出民航货邮周转量随时间推移而发展变化的动态规律，最后为更精确地预测月度民航货邮周转量，提出基于ARIMA-GM的组合预测模型，并对近几月民航货邮周转量进行较准确的短期预测，结果表明：组合模型能提高预测精度，在实际应用中ARIMA模型可用于非季节和季节的各类时间序列；灰色GM（1，1）模型能准确反映时间序列的增长趋势，两者相结合很好地解决了民航货邮周转量短期预测的实际问题，得到民航货邮周转量更精确的预测结论，能够对民航货邮市场的发展趋势进行宏观把握，有利于决策者的经济决策行为。

关键词: 民航货邮周转量, GM（1, 1）模型, 自回归移动平均（ARIMA）模型, 组合模型