[R-SIG-Finance] rugarch - warning

[alexios ghalanos]

Hi Harry,

This is nothing to worry about. The solver is visiting the default lower 
bounds for the GED shape which are admittedly too small (0.01). Until I 
change the defaults, simply change them as follows:
setbounds(gspec.ru)<-list(shape=c(0.1, 3))

Best,

Alexios

PS A new version (1.01-3) in the google code repository contains 
substantial improvements including exclusive use of xts (and returning 
xts objects) as well as a uniform set of extractor methods (sigma and 
fitted on almost all objects). Please try out/test and report any 
problems on the support page there.


On 02/02/2013 21:32, Harald Weiss wrote:
> Hi R-user group,
> I'm estimating GARCH models (GARCH and EGARCH) with the rugarch-package.
> As an error distribution I use the generalized error distribution.
> Performing a rolling
> estimation I receive for some days the following warning:
>
>     "In dged(x, nu = shape) : value out of range in 'gammafn'"
>
> The estimation results seem useful. I'm not sure whether I could use the
> results.
> I know that by using the solver "nloptr" the warning disappears
> (unfortunately,
> the warning occurs for other periods). I also know that the the gamma
> function in R
> is defined for x<=171 (perhaps this will explain the warning, but I
> don't understand
> how). So, my question is whether the results are feasible or not.
> Thanks,
> Harry
>
> My code is:
>
> DATA <- read.table("test-rugarch.csv",header=FALSE,sep=";")
> RDAX <- DATA[,1]
> D <- matrix(RDAX)
>
> for (i in 471:471) {
> ende <- 500+i
> start <- ende-499
> RDAXtmp <- D[start:ende,1]
> gspec.ru <- ugarchspec(variance.model=list(model="sGARCH",
> garchOrder=c(1,1)),
> mean.model=list(arfima=FALSE, armaOrder=c(0,0), include.mean=TRUE),
> distribution="ged")
> gfit <- ugarchfit(gspec.ru, RDAXtmp, fit.control = list(stationarity = 1))
> forc <- ugarchforecast(gfit, n.ahead=1)
> }
>
>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions should go.
>