摘要
GARCH模型是研究金融资产收益的重要模型,然而现有参数GARCH模型依然不能有效刻画金融资产收益偏态厚尾特性且存在模型设定风险。本文在非参数分布和GARCH模型基础上,建立半参数GARCH模型以提高模型的有效性;同时在贝叶斯框架内发展有效MCMC抽样解决模型的参数估计难问题,并利用DIC4研究模型比较问题;最后通过模拟研究和实证研究考察MCMC抽样的有效性,检验半参数GARCH模型在刻画金融资产收益特性和风险价值预测方面的实际效果。
GARCH model is an important model for describing financial asset returns.The emprical results show that GARCH model with one kind parametric distribution still can't capture the characteristics of returns,and it is hard to decide which distribuion is better.Therefore,this paper combines the nonparametric distribution and the GARCH model to build semiparametric GARCH model in order to improve the effectiveness of the model.This project proposes an effective MCMC algorithm to solve the problem of parameter estimation;and also proposes to use the modified DIC to study model selection.Lastly,we use simulation study and empricial study to analyze the effectiveness of the proposed MCMC sampling,also check the detailed ability of semiparametric GARCH model in decribing the characteristics of financial asset returns and forecasting the value at risk.
出处
《数理统计与管理》
CSSCI
北大核心
2015年第3期452-462,共11页
Journal of Applied Statistics and Management
基金
国家自然科学基金(11226221
71273048
11171065)
中国博士后基金(2013M540397)
江苏省高校自然科学基金(12KJB110009)
