Hi Rui,
I have tried different block values and below are the error message I got:
Error in as.POSIXlt.default(attributes(data)$times) :
do not know how to convert 'attributes(data)$times' to class ?POSIXlt?
out.list <- lapply(dat, gev, block = 10)
Error in optim(theta, negloglik, hessian = TRUE, ..., tmp = data) :
non-finite finite-difference value [2]
*From:* Rui Barradas <ruipbarradas at sapo.pt>
*To:* Roslina Zakaria <zroslina at yahoo.com>
*Cc:* 'r-help' <r-help at r-project.org>
*Sent:* Wednesday, July 17, 2013 8:49 PM
*Subject:* Re: [R] error message in gev
Hello,
It's better if you Cc the list, the odds of getting more and better
answers are greater.
Have you tried with different block values?
Rui Barradas
Em 17-07-2013 13:33, Roslina Zakaria escreveu:
> Hi Rui,
> Yes, I am using gev from 'evir'
> install.packages("evir")
> library(evir)
> I have tried your suggestion and follow are what I got.
>
>>out.list <- lapply(dat, gev, block = 20)
> Error in optim(theta, negloglik, hessian = TRUE, ..., tmp = data) :
> non-finite finite-difference value [1]
> In addition:Warning messages:
> 1: In FUN(X[[7L]], ...) : optimization may not have succeeded
> 2: In FUN(X[[7L]], ...) : optimization may not have succeeded
> 3: In sqrt(diag(varcov)) : NaNs produced
> 4: In FUN(X[[7L]], ...) : optimization may not have succeeded
> 5: In FUN(X[[7L]], ...) : optimization may not have succeeded
> 6: In FUN(X[[7L]], ...) : optimization may not have succeeded
> 7: In FUN(X[[7L]], ...) : optimization may not have succeeded
>
>
>
> Thank you.
>
>
> *From:* Rui Barradas <ruipbarradas at sapo.pt <mailto:ruipbarradas at sapo.pt>>
> *To:* Roslina Zakaria <zroslina at yahoo.com <mailto:zroslina at yahoo.com>>
> *Cc:* "r-help at r-project.org <mailto:r-help at r-project.org>"
<r-help at r-project.org <mailto:r-help at r-project.org>>
> *Sent:* Wednesday, July 17, 2013 5:51 PM
> *Subject:* Re: [R] error message in gev
>
> Hello,
>
> You should say which package is gev() coming from. I believe it's from
> evir, and what follows assumes that.
> The problem is that 'dat' is a data.frame, a special type of list. It is
> a list of vectors and gev is expecting just one vector. You could try
> instead, to fit generalized extreme value distributions to each of the
> vectors in the data.frame,
>
> out.list <- lapply(dat, gev, block = 20)
>
>
> to get a list of 10 objects of class gev, like the help page says.
>
> Hope this helps,
>
> Rui Barradas
>
> Em 17-07-2013 10:33, Roslina Zakaria escreveu:
> >
> > Hi r-users,
> >
> > I would like to use gev and my data (annual rainfall ) is as follows:
> >
> >> head(dat,20) A B C D E F G H I J
> > 1 45.1 41.5 58.5 50.1 46.0 49.1 37.7 49.1 59.8 54.0
> > 2 50.3 39.8 49.4 56.4 49.4 48.8 42.1 49.8 49.4 58.3
> > 3 41.7 39.3 44.6 39.1 35.7 41.5 40.8 40.8 38.5 45.6
> > 4 50.7 33.9 48.4 28.2 35.5 39.1 61.4 17.0 30.7 38.3
> > 5 39.3 30.6 46.9 23.8 25.8 17.1 37.9 40.1 35.4 30.1
> > 6 50.2 43.6 55.5 26.7 36.6 20.6 35.2 46.0 52.2 30.5
> > 7 46.0 37.0 55.3 36.3 51.9 50.6 43.2 44.9 38.2 45.4
> > 8 46.0 15.4 35.7 27.3 40.2 21.2 41.1 65.7 35.7 28.7
> > 9 52.0 36.5 49.6 55.0 38.3 32.0 43.8 43.8 50.3 36.5
> > 10 57.5 51.0 47.0 63.0 43.0 46.5 41.5 82.8 50.5 38.5
> > 11 71.0 48.0 59.5 48.5 47.0 57.5 55.8 64.5 60.0 43.4
> > 12 57.5 52.5 58.5 42.0 62.0 44.0 47.5 59.5 45.0 51.5
> > 13 47.0 38.0 49.9 50.5 55.0 57.5 51.5 54.5 45.0 49.7
> > 14 66.5 44.4 60.0 99.5 54.5 41.3 57.0 69.5 33.0 58.4
> > 15 61.5 43.6 68.5 50.7 48.0 50.5 50.5 58.8 53.2 47.0
> > 16 55.5 58.5 55.5 45.8 48.0 57.0 52.0 48.5 58.5 437.5
> > 17 96.7 96.7 96.7 45.5 65.1 46.5 51.0 96.7 61.3 310.5
> > 18 71.5 55.5 63.0 52.5 59.0 55.5 55.5 51.1 47.0 49.5
> > 19 71.5 42.5 64.0 56.5 61.0 54.0 46.5 53.7 43.0 44.0
> > 20 57.5 51.5 51.3 34.0 55.0 61.0 55.5 51.3 51.3 44.5
> >
> > When I apply the gev function, I have the following error message:
> >
> > out <- gev(dat, 20) Error in gev(dat, 20) : (list) object cannot be
> coerced to type 'double'
> >
> > Can anybody explain to me what does it means. I tried to read the
> package manual but could not understand about the block.
> >
> > Thank you so much for any help given.
> > [[alternative HTML version deleted]]
> >
> >
> >
> > ______________________________________________
> > R-help at r-project.org <mailto:R-help at r-project.org>
<mailto:R-help at r-project.org <mailto:R-help at r-project.org>> mailing list