中长期预测模型GMDH自回归两水平数据组选择研究及实证分析

Study Choosing Data of Two-Level Algorithm of the Model of GMDHAR Prediction and Application

经济系统中的某些变量需要预测月均值 ,一年 1 2月的预测区间较长 ,其准确性会受到影响。用GMDH两水平算法来扩大可预测范围 ,又涉及到数据组的选择 :用月均值数据与年均值数据同时建模 ,或是用月均值数据与季节均值数据同时建模 ,是提高预测准确性的关键。本文用理论证明和实证分析表明。GMDH自回归两水平方法的预测 。

We need to predict the average value of variable in the economic system,the prediction is usually biased because the predicting intervals is long.It is key to choose the difference to predict it with GMDHAR. the result that we researched theproblems of prediction with the theory and methods in example shows that:predicting result with GMDHAR by “month yeas” is better than that by “month season”.