[R] R crashes using quantreg censored estimation (method fcen)

roger koenker rkoenker at uiuc.edu
Thu Nov 29 14:23:06 CET 2007


1.  As the posting guide advises,  queries about packages should be  
sent in the first instance
to the package author/maintainer before posting to the list.

2.  As the posting guide also advises, a _minimal_ reproducible  
example is desirable not one
that claims that somewhere in the middle of 1,000,000 bootstrap  
replications something goes
wrong.

I'll try to take a look at this with valgrind, but I can't promise  
any immediate resolution.


url:    www.econ.uiuc.edu/~roger            Roger Koenker
email    rkoenker at uiuc.edu            Department of Economics
vox:     217-333-4558                University of Illinois
fax:       217-244-6678                Champaign, IL 61820


On Nov 29, 2007, at 6:32 AM, Malte Brockmann wrote:

> Thanks for your advice, Gavin. I hope to comply with the posting  
> guide this time.
>
> With "crashing" I meant that the R-session is simply terminated  
> without an error message (Vista, R2.6.1 and R2.5.1 with precompiled  
> packages, sessionInfo see below). I could reproduce the error on a  
> linux machine (R2.5.1-1 for which I have compiled the packages on  
> my own). On the linux machine I at least get the following error  
> message:
>
> *** caught segfault ***
> address (nil), cause 'unknown'
>
> Traceback:
>  1: .Fortran("xys", as.integer(m), as.integer(n), as.integer 
> (p),     as.integer(R), as.integer(m + 5), as.integer(p + 2),  
> as.double(x),     as.double(y), as.double(tau), as.double(tol),  
> flag = integer(R),     coef = double(p * R), resid = double(m),  
> integer(m), double((m +         5) * (p + 2)), double(m), as.integer 
> (1), sol = double((p +         2)), dsol = double(m), lsol =  
> as.integer(0), xx = double(m *         p), yy = double(m),  
> as.integer(s), PACKAGE = "quantreg")
>  2: boot.rq.xy(x, y, s, tau)
>  3: boot.rq(x[s, ], y[s], tau, ...)
>  4: summary.fcrq(model.powell, R = r)
>  5: summary(model.powell, R = 1e+06)
>
> I used the following command to generate the model:
> model.powell <- rq(y ~ x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 + x9,  
> method="fcen", data=data_test)
> summary(model.powell, R=1000000)
>
> I used 1000000 replications to make sure that the error is triggered
>
> I uploaded the data to www.esbf.de/quantreg/data_test.RData
>
>
>
>
>
>
> -----Ursprüngliche Nachricht-----
> Von: Gavin Simpson [mailto:gavin.simpson at ucl.ac.uk]
> Gesendet: Donnerstag, 29. November 2007 11:45
> An: Malte Brockmann
> Cc: r-help at stat.math.ethz.ch
> Betreff: Re: [R] R crashes using quantreg censored estimation  
> (method fcen)
>
> On Thu, 2007-11-29 at 09:09 +0100, Malte Brockmann wrote:
>> Dear List,
>>
>> my R session crashes if I repeatedly run the summary function on a
>> quantreg object generated using the censored estimation method or  
>> if I
>> set the number of bootstrap repititions reasonably high. Are there  
>> any
>> known incompabilities between quantreg and Windows Vista & AMD64? The
>> error is reproducable using R 2.6.1 and R 2.5.1.
>
> Err, what does this have to do with Jared's reply to Max's posting  
> about
> suitable textbooks and R?
>
> Please don't jump on a thread to ask your own question. Start your own
> thread.
>
> If you want people to help you with your problem, you are probably  
> going
> to have to give a reproducible example - "my R session crashes" is way
> below what is requested in the posting guide. Even the statements  
> about
> what you did are not enough - people need to be able to reproduce your
> problem on their machines.
>
> Read the posting guide (URL in the footer of this email) and resend a
> message to the list - a new one mind, so you start a new thread. And
> make sure you include a reproducible example. If you can't send your
> data, try it with one of the data sets that come with quantreg and  
> if it
> fails with that also, use that data set in your example.
>
> You should also consider contacting the maintainer of quantreg.
>
> HTH
>
> G
>
>>
>>> sessionInfo()
>> R version 2.6.1 (2007-11-26)
>> i386-pc-mingw32
>>
>> locale:
>> LC_COLLATE=German_Germany.1252;LC_CTYPE=German_Germany. 
>> 1252;LC_MONETARY=German_Germany. 
>> 1252;LC_NUMERIC=C;LC_TIME=German_Germany.1252
>>
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods
>> base
>>
>> other attached packages:
>> [1] quantreg_4.10 SparseM_0.74
>>
>> Thanks,
>> Malte
>> ______________________________________________
>> 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.
>>
>> ______________________________________________
>> 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.
> -- 
> %~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
>  Dr. Gavin Simpson             [t] +44 (0)20 7679 0522
>  ECRC, UCL Geography,          [f] +44 (0)20 7679 0565
>  Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
>  Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/
>  UK. WC1E 6BT.                 [w] http://www.freshwaters.org.uk
> %~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
>
> ______________________________________________
> 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.



More information about the R-help mailing list