[R] Opaque error message and R GUI crashing while using glmnet

[Duncan Murdoch]

On 24/02/2010 1:18 AM, Katherine Fitzgerald wrote:
> Hello all,
> 
> I am having trouble with the glmnet package.  I used it for the first
> time last week, after updating R to the current version (2.10.1,
> running in Windows XP).  It seemed to be working fine as I explored
> what it could do with one of my datasets.  A day or two later, using
> the exact same dataset, it suddenly stopped working.

This sounds like a bug in glmnet.  Have you tried contacting it's 
maintainer?

Duncan Murdoch

> 
> I get the following error every time I try to use glmnet with real Y
> values from my dataset:
> "Error: in glmnet fortran code - Null probability for class 1 < 1.0e-5"
> 
> Once I get that error, after I type a few more commands, I often get
> the following pop-up, accompanied by a button to send an error report
> to Microsoft:
> "R for Windows GUI front-end has encountered a problem and needs to
> close.  We are sorry for the inconvenience."
> When this happens the R window abruptly closes.  It does not appear to
> matter what the commands are following the error message.  It could be
> something like fix(Y) or help(glmnet).
> 
> I have tried different real datasets, as well as fake datasets like in
> the example at the end of the glmnet help file.  The error occurs
> whenever I use my real Y data, but not when I use randomly produced
> fake Y data.  It does not appear to matter whether the X values are
> real or made-up.
> 
> I have tried uninstalling and reinstalling R.  I've also tried using
> the latest patched version of R, as well as earlier versions.  While
> using version 2.9.2, after entering the command:
>> help(glmnet)
> I got the following pop-up message:
> "Microsoft Visual C++ Runtime Library
> Runtime Error!
> Program: C:\Program Files\R\R-2.9.2\bin\Rgui.exe
> This application has requested the Runtime to terminate it in an
> unusual way.  Please contact the application's support team for more
> information."
> R did not close down after this but was hung up with the hourglass mouse symbol.
> 
> I've attached a text file with some Y values in it.  I've been able to
> reproduce the error message using the following commands:
> 
> Y = as.matrix(read.table("C:\\folderpath\\Y_example.txt", header=T))
> X=matrix(rnorm(100*20),100,20)
> example = glmnet(X,Y,family="binomial")
> 
> If anyone can explain what this error means, or why glmnet would
> suddently stop working, I would appreciate your insight!
> 
> Thanks,
> 
> 
> 
> ------------------------------------------------------------------------
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.