On Jun 18, 2015, at 7:11 PM, alexios <alexios at 4dscape.com> wrote:
Hi Daniel,
That's a bug...I've upload a patch to my development repo. You can install as follows:
library(devtools)
install_bitbucket("alexiosg/rmgarch")
This should now work:
library(rmgarch)
data(dji30ret)
library(xts)
dj<-xts(dji30ret, as.Date(rownames(dji30ret)))
spec<-gogarchspec(mean.model=list(model="AR"))
scen<-fscenario(dj[,1:10], sim=1000, roll=1, model="gogarch", spec = spec, n.comp=3, gfun="tanh", solver.control=list(trace=1))
dim(fitted(scen))
[1] 1000 10 2
Best,
Alexios
On 18/06/2015 23:10, daniel melendez via R-SIG-Finance wrote:
Hello - I selected to keep 4 of the 10 components associated with my series for modeling purposes. However, when trying to generate scenarios using fscenario I get a warning as follows: Error in tail(preresiduals, mx) %*% t(A) : non-conformable arguments
In addition: Warning messages:
1: In matrix(fit at mfit$Y[(T - p + i):(T - 1 + i), , drop = FALSE], ncol = m, :
data length [4] is not a sub-multiple or multiple of the number of columns [10]
2: In matrix(filt at mfilter$factor.sigmas[(T - p + i):(T - 1 + i), , :
data length [4] is not a sub-multiple or multiple of the number of columns [10] Any help would be greatly appreciated. Does this message mean that I must use all 10 components for generating a scenario? Regards
DA Melendez
[[alternative HTML version deleted]]