R help - Aug 2007 - lasso/lars error

I'm having the exact problem outlined in a previous post from 2005 -
unfortunately the post was never answered:

http://tolstoy.newcastle.edu.au/R/help/05/10/15055.html

When running:

lm2=lars(x2,y,type="lasso",use.Gram=F)

I get an error:

Error in if (zmin < gamhat) { : missing value where TRUE/FALSE needed

...when running lasso via lars() on a 67x3795 set of predictors.  I hacked
the lars() function to print out the zmin, gamhat, and also z1 values (used
to calculate zmin) - it seems that the error comes up during the lasso
process when all the z1's are negative - unfortunately I have no idea what
this means. This leads to both zmin and gamhat being NA.  The error can be
worked around by forcing max.steps to some low value, but this isn't an
ideal solution.

I hope this post won't suffer the same fate as the previous one on the same
topic.  If I can provide any more useful information please let me know.

Thanks,

Jake
> sessionInfo()
R version 2.5.0 (2007-04-23)
i686-pc-linux-gnu

locale:
LC_CTYPE=en_US.UTF-8;LC_NUMERIC=C;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=en_US.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=C

attached base packages:
[1] "stats"     "graphics"  "grDevices"
"utils"     "datasets"  "methods"
[7] "base"

other attached packages:
        lars randomForest      corpcor  cairoDevice
     "0.9-7"     "4.5-18"      "1.4.5"       
"2.3"

	[[alternative HTML version deleted]]

This looks like a question for the package maintainer (as the posting 
guide said), as was the one you referred to.  Please bear in mind
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
Anyone who is going to be able to help will need such an example.

It seems that a reply like this needs to go on the archive for future 
reference.


On Thu, 2 Aug 2007, Jake Michaelson wrote:
> I'm having the exact problem outlined in a previous post from 2005 -
> unfortunately the post was never answered:
>
> http://tolstoy.newcastle.edu.au/R/help/05/10/15055.html
>
> When running:
>
> lm2=lars(x2,y,type="lasso",use.Gram=F)
>
> I get an error:
>
> Error in if (zmin < gamhat) { : missing value where TRUE/FALSE needed
>
> ...when running lasso via lars() on a 67x3795 set of predictors.  I hacked
> the lars() function to print out the zmin, gamhat, and also z1 values (used
> to calculate zmin) - it seems that the error comes up during the lasso
> process when all the z1's are negative - unfortunately I have no idea
what
> this means. This leads to both zmin and gamhat being NA.  The error can be
> worked around by forcing max.steps to some low value, but this isn't an
> ideal solution.
>
> I hope this post won't suffer the same fate as the previous one on the
same
> topic.  If I can provide any more useful information please let me know.
>
> Thanks,
>
> Jake
>
>> sessionInfo()
> R version 2.5.0 (2007-04-23)
> i686-pc-linux-gnu
>
> locale:
>
LC_CTYPE=en_US.UTF-8;LC_NUMERIC=C;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=en_US.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=C
>
> attached base packages:
> [1] "stats"     "graphics"  "grDevices"
"utils"     "datasets"  "methods"
> [7] "base"
>
> other attached packages:
>        lars randomForest      corpcor  cairoDevice
>     "0.9-7"     "4.5-18"      "1.4.5"       
"2.3"
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at stat.math.ethz.ch 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.
>
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

I'm having the same error too, but it's very rare though... (especially
when
using datasets with many variables, say 25). A somewhat odd, but seemingly
effective, way to solve this, is by writing try(lars(x,y),silent = T). The
try-command in this shape will not generate an error on the console and
therefore, any running programs are not aborted. When there is no error, you
simply get the LARS object. When there is an error, you get a character
string (can be checked by 'is.character()'). I suggest to perturb the
data a
little to get it to work...

Best Regards,

Timmie Smet

B.A.P. Operations Research and Business Statistics
University of Leuven, Belgium
-- 
View this message in context:
http://r.789695.n4.nabble.com/lasso-lars-error-tp831400p3173480.html
Sent from the R help mailing list archive at Nabble.com.