基于权函数的GARCH模型稳健建模及实证分析  

Robust Modling of GARCH Model Based on Weighted Function and Empirical Analysis

作　　者：苏拥英[1] 俞健[2] 

机构地区：[1]广州科技贸易职业学院,广东广州511442 [2]广州大学数学与信息科学学院,广东广州510006

出　　处：《数学的实践与认识》2015年第18期28-36,共9页Mathematics in Practice and Theory

基　　金：广州市教育科学"十二五"规划2014年面上项目(1201431134);广州市属高校2012年度立项课题(2012B156)

摘　　要：异常值的存在会对时间序列波动率模型的识别及参数估计会产生重要影响,采用Tukey双权法权函数对被拟合相关序列模型的残差进行变换,再将变换后的残差序列对波动率模型进行稳健识别及建模,模拟及实证分析表明稳健识别及估计方法具有很好的耐抗性,而且能更好的捕捉到资产收益率的波动性.The outlier has an important impact on the identification and estimate of time Series volatility model,this paper adopts Tukey double weighted function to transform the residuals of fitted correlation sequence model,and then the robust identification and estimate simulation and empirical analysis of the residual sequence transformed on volatility models show that robust identification and estimate method in this paper has good resistance,and can better capture the volatility of return on assets.

关 键 词：权函数 GARCH模型 稳健建模 收益率 R语言 

分 类 号：O211.61[理学—概率论与数理统计]