哈哈既然是交作业,何必为难自己
选一个Extreme Values这么大的数据
模型建立后,就有残差及fitted values,不是预测
> ans[1:5,]
data fitted residuals sigma
1 -0.008298803 -0.0082988028 0.000000000 0.011257653
2 0.000000000 -0.0059713111 0.005971311 0.010149800
3 0.000000000 -0.0025937958 0.002593796 0.009544355
4 -0.008368250 -0.0008729045 -0.007495345 0.008705944
5 0.000000000 -0.0022060523 0.002206052 0.008564666
要预测的是 sigma & series
> f
*------------------------------------*
* GARCH Model Forecast *
*------------------------------------*
Model: iGARCH
Horizon: 10
Roll Steps: 0
Out of Sample: 0
0-roll forecast:
sigma series
1192 0.008360 0.002483
1193 0.008447 0.002370
1194 0.008532 0.002282
1195 0.008617 0.002214
1196 0.008701 0.002161
1197 0.008785 0.002121
1198 0.008867 0.002089
1199 0.008949 0.002065
1200 0.009030 0.002046
1201 0.009110 0.002032
###################
library(rgarch)
ppiaco=read.csv("ppiaco.csv")
ppiaco=ppiaco[,1]
n=length(ppiaco)
# calculate log-returns for GARCH analysis
p.ret=log(ppiaco[2:n])-log(ppiaco[1:(n-1)])
variance.model=list(model="iGARCH",garchOrder=c(1,1))
mean.model=list(armaOrder=c(1,1),include.mean=T,garchInMean=F)
spec=ugarchspec(variance.model=variance.model,mean.mode=mean.model,distribution.model="sged")
fit=ugarchfit(data=p.ret,spec=spec)
fit
f=ugarchforecast(fit, n.ahead=10)
f
plot(fit,which=9)
ans=as.data.frame(fit)
ans[1:5,]