基于GMDH的VAR滞后选择方法及实证分析

A New Method of Choosing Variable Lag in the Model of VRA and its Application

在VAR的动态模型中 ,滞后数的选择是成功的关键。对于复杂系统建模 ,根据GMDH理论的有关数据 ,用计算机合理地调协变量与滞后变量之间的关系 ,确定模型的最优动态结构和最优滞后数 ,以提高预测的准确性。给出了VAR的传统方法与GMDH建模方法 ,应用经济变量进行实证分析 ,阐明了GMDH算法的优势在于比传统的方法更完善 ,预测效果更准确。

Based on the principle of self organization,the multi level algorithm model is a system with multilayer algorithm It mainly solves the problems of choosing variable lag in the model of VRA of complex system modeling In this article,the advantages of the proposed new method for the model of VAR to choose variable lag and its feasibility are demonstrated with examples