[R] Optim error: initial value in 'vmmin' is not finite

Ben Bolker bolker at ufl.edu
Thu Jan 29 23:20:53 CET 2009


SnowManPaddington <wiwiana <at> gmail.com> writes:

> 
> 
> Error in optim(method = "BFGS", c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  : 
>   initial value in 'vmmin' is not finite
> 
> I am running a logit model with latent class segments. I successfully got
> estimates for 2 segments. However when I tried to increase the no. of
> segments, I got this error message at the end. I checked my code again but
> can't find anything wrong. Is this error message related to something
> different from the code? Is that anything to do with the initial value? what
> does vmmin mean exactly?
> 

   As the error message says, there is a problem computing
the initial values for the model fitting procedure. These
initial values may be hidden within the code you are using.

  Please read the posting guide -- we need a lot more information
before we can help you.  If at all possible provide a reproducible
example, and tell us what packages/functions you are using (and
the output of sessionInfo() ).

  good luck
   Ben Bolker




More information about the R-help mailing list