R help - Jan 2010 - Problems with fitdistr

Hi,
I want to estimate parameters of weibull distribution. For this, I am using
fitdistr() function  in MASS package.But when I give
fitdistr(c,"weibull") I
get a Error as follows:- 
     Error in optim(x = c(4L, 41L, 20L, 6L, 12L, 6L, 7L, 13L, 2L, 8L, 22L, 
: 
     non-finite value supplied by optim
Any help or suggestions are most welcomed
-- 
View this message in context:
http://n4.nabble.com/Problems-with-fitdistr-tp1334772p1334772.html
Sent from the R help mailing list archive at Nabble.com.

Try to pass a start value to help optim (see ?fitdistr)
Ciao!
	mario

vikrant wrote:
> Hi,
> I want to estimate parameters of weibull distribution. For this, I am using
> fitdistr() function  in MASS package.But when I give
fitdistr(c,"weibull") I
> get a Error as follows:- 
>      Error in optim(x = c(4L, 41L, 20L, 6L, 12L, 6L, 7L, 13L, 2L, 8L, 22L, 
> : 
>      non-finite value supplied by optim
> Any help or suggestions are most welcomed
-- 
Ing. Mario Valle
Data Analysis and Visualization Group            | http://www.cscs.ch/~mvalle
Swiss National Supercomputing Centre (CSCS)      | Tel:  +41 (91) 610.82.60
v. Cantonale Galleria 2, 6928 Manno, Switzerland | Fax:  +41 (91) 610.82.82

Do you have any zeros in your data?
fitdistr() will need start values (see the code),
but even with start values, optim() will have problems.

x <- rweibull(100, 2, 10)
fitdistr(x, "weibull")  ## no problem
fitdistr(c(0,x), "weibull")  ## your error message
fitdistr(c(0,x), "weibull",start=list(shape=2,scale=10))
   ## still an error message from optim()

  -Peter Ehlers

vikrant wrote:
> Hi,
> I want to estimate parameters of weibull distribution. For this, I am using
> fitdistr() function  in MASS package.But when I give
fitdistr(c,"weibull") I
> get a Error as follows:- 
>      Error in optim(x = c(4L, 41L, 20L, 6L, 12L, 6L, 7L, 13L, 2L, 8L, 22L, 
> : 
>      non-finite value supplied by optim
> Any help or suggestions are most welcomed
-- 
Peter Ehlers
University of Calgary

vikrant wrote:
> Hi,
> I want to estimate parameters of weibull distribution. For this, I am using
> fitdistr() function  in MASS package.But when I give
fitdistr(c,"weibull") I
> get a Error as follows:- 
>      Error in optim(x = c(4L, 41L, 20L, 6L, 12L, 6L, 7L, 13L, 2L, 8L, 22L, 
> : 
>      non-finite value supplied by optim
> Any help or suggestions are most welcomed
Use function pelwei() in package lmom.


J. R. M. Hosking

Yes I do have zeros in my data. But I m not able to understand y inclusion of
zeros results in error messages, because range for x in weibull distribution
is x>=0. Can you please clarify this doubt? 
-- 
View this message in context:
http://n4.nabble.com/Problems-with-fitdistr-tp1334772p1401513.html
Sent from the R help mailing list archive at Nabble.com.