ARCH模型的诊断分析

Diagnose analysis in ARCH models

探讨了 ARCH模型的诊断分析和变结构建模问题 .所提出的分整增广 GARCH- M模型包括了几乎所有现有的 ARCH模型 ,基于分整增广 GARCH- M模型克服了 ARCH模型在模型设定检验 ,长记忆检验和参数估计等方面的障碍 .利用分段建模方法来检测模型结构变化点和分段变化模型的选择 .最后 ,以上海证券综合事业股票指数为数据 。

The diagnose analysis and modeling problem with structure change in ARCH models are discussed in this paper.The paper proposses fractional integrated augmented GARCH M model which can embrace almost all the current ARCH models. Based on the fractional integrated augmented GARCH M model, the handicap in model specification test, long memory test and parameters estimation is oversome.The segment modeling method is used to detect the structure changing points of the ARCH model and test the models for segment changing models.Finally, with the composite index of closing price in Shanghai stock market, it is shown to illustrate the effectiveness of the diagnose analysis.