1

I would like to understand the issue with plotting a fitdistrplus object:

    ftnb = tryCatch(expr = fitdist(unlist(x[,ftr]),distr="nbinom",discrete = TRUE),
                      error = function(err) {})
if (class(ftnb) %in% c("fitdist")) {
  print(summary(ftnb))
  plot(ftnb))
}

The summary of my distribution fit is as follows:

Fitting of the distribution ' nbinom ' by maximum likelihood Parameters : 

         estimate Std. Error
size 5.915112e-01         NA
mu   1.330646e+11         NA
Loglikelihood:  -2543.973   AIC:  5091.946   BIC:  5097.074 
Correlation matrix:
[1] NA

I clearly do not see a fit, as I see an "NA" in the standard errors of the 2 parameter estimates. The plot command comes up with the following error, which I would like to understand:

Error in seq.default(min(xval), max(xval), by = 1) :   'by' argument is much too small

Please help.

Srivatsa Srinath
  • 177
  • 2
  • 10
  • 1
    @MLavoie means http://stats.stackexchange.com/questions/193741/r-fitdistrplus-seek-help-in-understanding-na-in-std-error –  Feb 03 '16 at 00:54
  • @Pascal ... I have posted a different question in stats.stackexhange, as I wasn;'t sure if it is an R issue or statically there is an issue as well. As I had mentioned, I will add the data relevant to the question posed. – Srivatsa Srinath Feb 03 '16 at 08:34

0 Answers0