高低频预测模型结合的最优化方法

Optimization method for merging the short and long run forecasting models

结合用于预测同一经济变量的高低频预测模型能够集中两类模型的优点 ,使得低频预测结果动态化 ,是同时具有理论意义和实用价值的课题 .本文用最优化方法对此进行了研究 .文中提出了高频序列有补充信息时对低频预测结果的动态调整方法 .该方法已用于中国进口与关税税收预测支持系统的研制 。

Merging the short and long run forecasting models used for an economic variable can combine the advantages of these models and obtain the dynamic forecasting results of long run model.Therefore,it is interesting in both theory and practice.This paper solves the problem by means of optimization method.A dynamic adjusting method is presented when some additional information in the high frequency data is available.This method is used in developing the China Import and Tariff Revenue Forecasting Support System.The testing results showed that the new method is of good forecasting effects and practicable.