Message-ID: <6rof35gkm8.fsf@bates4.stat.wisc.edu>
Date: 2003-04-17T16:21:51Z
From: Douglas Bates
Subject: glm: NA/NaN/Inf in foreign function call
In-Reply-To: <F934BF2710426B44833B8677AF47F2467486D3@mail3.tay.ac.uk>
J.Illian at abertay.ac.uk writes:
> Dear all,
> I am running the glm function (family: Poisson) in R and most of the time it
> works fine. But sometimes, with large data sets, I get the following error
> message:
> Error: NA/NaN/Inf in foreign function call (arg 1)
>
> What puzzles me most about this is the fact that there are no missing values
> in my vectors at all- so where do the missing values come from? Could this
> be related to the size of the vectors I use?
Try using control = list(trace = TRUE) in your call to glm to see what
the progress of the IRLS algorithm is.