基于ARFIMA-GARCH模型的混成检验  

PORTMANTEAU TEST BASED ON THE ARFIMA-GARCH MODEL

作　　者：玄海燕[1] 史永侠 张玉春[1] 徐广业[1] 

机构地区：[1]兰州理工大学经济管理学院,甘肃兰州730050 [2]兰州理工大学理学院,甘肃兰州730050

出　　处：《数学杂志》2017年第3期474-480,共7页Journal of Mathematics

基　　金：国家自然科学基金资助(11261031)

摘　　要：本文研究了ARFIMA-GARCH模型的混成检验问题.基于拟极大指数似然估计,给出了平方残差自相关函数的渐近性,进而建立了基于平方残差自相关函数的混成检验统计量.通过实例分析,表明可利用基于平方残差自相关函数的混成检验统计量来诊断检验由拟极大指数似然估计方法拟合的ARFIMA-GARCH模型.In this paper, we study the portmanteau test problem of ARFIMA-GARCH model. Based on the quasi-maximum exponential likelihood estimator, the asymptotic of squared residual autocorrelation function is given, and the portmanteau test statistic based on squared residual autocorrelation function is established. By the analysis of a real example, it is showed that we can use the portmanteau test statistic based on squared residual autocorrelation function to the diagnostic test of ARFIMA-GARCH model fitting by quasi-maximum exponential likelihood estimator.

关 键 词：ARFIMA-GARCH模型 混成检验 拟极大指数似然估计 

分 类 号：O212.1[理学—概率论与数理统计]