马尔可夫域变模型在我国季度GDP增长率序列建模中的应用

The Markov Regimes Switching Model of Quarterly GDP Growth Ratio Series in China

本文建立了我国季度GDP同比增长率序列的马尔可夫域变模型,通过与线性AR(X)、LSTAR和ARCH等模型的比较,结果表明它更好地刻画了研究对象的均值、波动性和动态结构存在域变行为的非线性特征.

This paper examines the nonlinear characteristic of the quarterly GDP growth ratio series in China using Markov regimes switching model. The result demonstrates it describes the time - varying behavior of the series more appropriately than linear AR（X） model, LSTAR model or ARCH model.