R help - Jun 2010 - glmnet strange error message

Hello fellow R users,

I have been getting a strange error message when using the cv.glmnet
function in the glmnet package. I am attempting to fit a multinomial
regression using the lasso. covars is a matrix with 80 rows and roughly 4000
columns, all the covariates are binary. resp is an eight level factor. I can
fit the model with no errors but when I try to cross-validate after about 30
seconds I get the following:

> glmnet.fit = glmnet(covars,resp,family="multinomial")
> glmnet.cv =
cv.glmnet(covars,resp,family="multinomial",type="class")
Error in if (outlist$msg != "Unknown error") return(outlist) : 
  argument is of length zero

It seems like it makes it through the first couple folds but trips up
somewhere in the middle.
The example in the documentation works perfectly on my machine. Any ideas on
what the problem may be?

Thanks!
Dave
-- 
View this message in context:
http://r.789695.n4.nabble.com/glmnet-strange-error-message-tp2240458p2240458.html
Sent from the R help mailing list archive at Nabble.com.

Could you give us the traceback? (In case you don't know, just type
traceback() right after you got the error message.) I can't reproduce the
error, so it gets a bit difficult to solve without having the real data.

Cheers
Joris

On Wed, Jun 2, 2010 at 6:51 PM, Dave_F <friedenbergd@battelle.org> wrote:
>
> Hello fellow R users,
>
> I have been getting a strange error message when using the cv.glmnet
> function in the glmnet package. I am attempting to fit a multinomial
> regression using the lasso. covars is a matrix with 80 rows and roughly
> 4000
> columns, all the covariates are binary. resp is an eight level factor. I
> can
> fit the model with no errors but when I try to cross-validate after about
> 30
> seconds I get the following:
>
>
> > glmnet.fit = glmnet(covars,resp,family="multinomial")
> > glmnet.cv =
cv.glmnet(covars,resp,family="multinomial",type="class")
> Error in if (outlist$msg != "Unknown error") return(outlist) :
>  argument is of length zero
>
> It seems like it makes it through the first couple folds but trips up
> somewhere in the middle.
> The example in the documentation works perfectly on my machine. Any ideas
> on
> what the problem may be?
>
> Thanks!
> Dave
> --
> View this message in context:
>
http://r.789695.n4.nabble.com/glmnet-strange-error-message-tp2240458p2240458.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help@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.
>


-- 
Joris Meys
Statistical Consultant

Ghent University
Faculty of Bioscience Engineering
Department of Applied mathematics, biometrics and process control

Coupure Links 653
B-9000 Gent

tel : +32 9 264 59 87
Joris.Meys@Ugent.be
-------------------------------
Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php

	[[alternative HTML version deleted]]

I think I have figured out the problem. The response variable has one level
with only one observation -- hence when that one observation was in the test
set the training data did not contain any observations from that level and
so did not fit a logistic model for that level causing the somewhat cryptic
error message.


-- 
View this message in context:
http://r.789695.n4.nabble.com/glmnet-strange-error-message-tp2240458p2241566.html
Sent from the R help mailing list archive at Nabble.com.