Displaying 1 result from an estimated 1 matches for "llh1".
Did you mean:
llh
2011 Feb 26
0
A problem about realized garch model
...{
logh[i] = omega+bet*logh[(i-1)]+gam*logx[(i-1)]
u[i] =
logx[i]-xi-phi*logh[i]-tau1*(r[i]/sqrt(exp(logh[i])))-tau2*((r[i]/sqrt(exp(logh[i])))^2-1)
}
h = exp(logh)
hh = sqrt(h)
llh = -sum(log(garchDist(r, hh)))-sum(log(measureDist(u,sigma.u)))
llh
}
fitting<-function(data)
{
LLH1<-function(theta) {LLH(data,theta)}
ini = c(0.1,0.6,0.3,0.07,1,-0.03,0.1,0.4)
for(i in 1:10)
{
fit = optim(ini,LLH1,control=list(trace=2),hessian=T)
ini = fit$par
}
hessian = fit$hessian
std = sqrt(diag(solve(hessian)))
A = rbind(fit$par,std)
A
}
=====================
Does anyone can...