基于波动门限效应GARCH模型的波动率预测  

The Volatility Forecast Based on the Fluctuation Threshold Effect GARCH Model

作　　者：王沁[1] 乔高秀[1] 

机构地区：[1]西南交通大学数学学院,四川成都610031

出　　处：《武汉理工大学学报（信息与管理工程版）》2017年第5期575-580,共6页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：国家自然科学基金项目(71371157);教育部人文社会科学研究青年基金项目(17YJC790119)

摘　　要：基于大的波动的稳定中心不同于小的波动的稳定中心的问题,提出一类具有波动门限效应的GARCH模型。研究了波动门限效应GARCH模型的参数估计,并通过AIC最小原则选择了门限阈值。根据MAE、RMSE和MAPE这3个损失函数,与传统GARCH模型的预测能力进行了比较分析。实证结果表明,波动门限效应GARCH模型在波动性预测方面比传统GARCH模型的效果更好。Based on the stability center of a large fluctuation is different from that of a small one,a special GARCH model with fluctuation threshold effect is proposed. How to estimate these parameters in the fluctuation threshold effect GARCH model is researched. The threshold value is selected by the minimum principle of the AIC value. The forecasting ability is compared between the traditional GARCH model and the fluctuation threshold GARCH model through three loss function（ MAE,RMSE and MAPE）. The results show that the fluctuation threshold effect GARCH model is better than the traditional GARCH model in terms of volatility forecasting.

关 键 词：波动门限效应 GARCH模型 AIC值 门限阈值 波动预测 损失函数 

分 类 号：F064.1[经济管理—政治经济学]