假设一定要在R实现的话,

请用package(rgarch).

这个结果跟在E-views是接近的.

variance.model=list(model="fGARCH",garchOrder=c(1,1),submodel="GARCH" );
mean.model=list(armaOrder=c(3,0),include.mean=T,garchInMean=F,inMeanType=1, fixed.pars=list(ar1=0,ar2=0)) ;
spec=ugarchspec(variance.model=variance.model,mean.mode=mean.model,distribution.model="std");
fit=ugarchfit(data=data,spec=spec,out.sample=0)
fit

Optimal Parameters
--------------------------
        Estimate  Std. Error  t value Pr(>|t|)
mu      0.031234    0.028875  1.08170 0.279387
ar1     0.012873    0.019377  0.66434 0.506474
ar2    -0.001446    0.003798 -0.38073 0.703405
ar3     0.056283    0.019590  2.87308 0.004065
omega   0.040826    0.013945  2.92760 0.003416
alpha1  0.078060    0.013820  5.64815 0.000000
beta1   0.912741    0.014631 62.38241 0.000000
shape   5.014735    0.549575  9.12476 0.000000

###################e-views   
Coefficient Std. Error z-Statistic Prob.  
   
AR(3) 0.056202 0.019977 2.813417 0.0049
   
Variance Equation   
   
C           0.037720 0.012220 3.086726 0.0020
RESID(-1)^2 0.075827 0.011784 6.434532 0.0000
GARCH(-1)   0.915578 0.012281 74.55062 0.0000
   
T-DIST. DOF 5.168727 0.591498 8.738364 0.0000

看过了

mean.model有这个功能

variance.model似乎没这个功能

上传rgarch-methods.R供你参考

rgarch-methods.rar
大小:(13.32 KB)

 马上下载

本附件包括：

• rgarch-methods.R