我国航空客运量需求预测模型:基于随机前沿预测模型和模型平均

针对多数研究以产品或服务的历史消费量来代替不可观测的需求量而导致的需求预测出现实质性偏差的问题，本文将包含技术无效率项的随机前沿预测模型应用于航空客运量需求的预测，从而有效解决实质性偏差的问题.同时我们在此基础上引入一种模型平均权重确定方法，即通过最小化M折交叉验证准则（CV_M）确定候选模型权重.本文证明了该方法在理论上的最优性.由于模型中技术无效项的存在，我们可以同时预测航空客运量的实际产生量和需求量，实证研究也表明，相比其他常用的预测方法，该方法在预测航空客运量中长期的实际产生量上更具优势.

A demand forecasting model for air passenger traffic in China: Based on stochastic frontier analysis model and model averaging

Considering that the historical consumption quantities of a product or service instead of the unobservable demands will result in the substantial forecasting error, we propose to apply the stochastic frontier analysis model including the inefficiency of producer into the air travel demand forecasting to solve this problem. In this study, we introduce a model averaging weight choice method by minimizing an M-fold cross-validation criterion (CV_M). The resulting model averaging estimator is asymptotically optimal. Due to existing the inefficiency of producer in model, we can both predict the demand and actual passenger traffic. Meanwhile, an empirical application of air travel demand forecasting is implemented. The result shows our method outperforms other common methods in terms of the medium and long actual passenger traffic forecasting.