R help - Jun 2010 - fgev + negative location parameter for positive data?

Hi R,

 

I have a vector with all positive values and have fit the GEV
distribution (shape parameter=0) to the data and I get the negative
value of the location parameter. Is this possible?

 

library(evd)

r=read.table("clipboard",header=F)[,1]
> length(r)
[1] 2087
> sum(r>=0) #All data points are positive
[1] 2087
> sum(r>=1000000) #All data points are positive
[1] 2087

 

 

Now,
> fgev_fit=fgev(r,std.err = FALSE)
> fgev_fit
 

Call: fgev(x = r, std.err = FALSE) 

Deviance: 82397.2 

 

Estimates

       loc       scale       shape  

-1.977e+07   1.141e+08   1.555e-02  

 

Optimization Information

  Convergence: successful 

  Function Evaluations: 30 

  Gradient Evaluations: 5

 

 

When all the values in the data are positive, how can the location
parameter of the fit be negative? Or is it something to do with the
large deviation? Please let me know, whether my assumptions are correct.

 

 

 

Thanks and Regards,

Shubha Karanth 

 

This e-mail may contain confidential and/or privileged i...{{dropped:13}}