# [R] Error while fitting gumbel copula

Martin Maechler maechler at stat.math.ethz.ch
Wed Aug 10 18:08:03 CEST 2016

```>>>>> Isaudin Ismail <isaudin at gmail.com>
>>>>>     on Tue, 9 Aug 2016 14:30:18 +0100 writes:

> Dear R experts,
> I have 5 time series of data (A, B, C, D and E) with all same lengths. All
> series exhibit gamma distribution except for B which is lognormal
> distribution. I am using copula package to model the joint-distribution of
> these 5 times series.

> I have selected Archimedean copula and successfully fitted  Frank and
> Clayton copula. The problem is when trying to fit Gumbel copula.

> The following are the codes I used to run in R.

> # Data of 5 time series

> A <- A
> B <- B
> C <- C
> D <- D
> E <- E

well, the above is really an "interesting" block of R code ;-)

--

More seriously, please learn to use reproducible examples,
e.g., from here
http://bit.ly/MRE_R (nice to remember: MRE = Minimal Reproducible Example)
or here

notably I as maintainer of the package 'copula' which you are
using (without saying so).

With regards,
Martin Maechler

> # Combined between A and C
> A+C <- A + C

> gumbel.copula <- gumbelCopula(dim = 5)
> m <- pobs(as.matrix(cbind(A+C, B, D, E)))
> fit.gumbel<- fitCopula(gumbel.copula, m, method = 'ml')

> And the error while trying to fit gumbel copula:

> Error in optim(start, loglikCopula, lower = lower, upper = upper, method =
> method,  :
> non-finite finite-difference value [1]
> In .local(copula, tau, ...) : tau is out of the range [0, 1]

> Appreciate all help!

> Many thanks,
> Isaudin

> [[alternative HTML version deleted]]

> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help