Hi there,
This problem has been dogging me for a bit, and I'm trying to
figure out why. When running the the subsets function in the leaps
library, R is giving me the following error message
> lvodsub <- subsets(pred, resp$LVOD)
Warning message:
XHAUST returned error code -999 in: leaps.exhaustive(a, really.big really.big)
but this still happens if I add the really.big option:
> lvodsub <- subsets(pred, resp$LVOD, really.big=T)
Warning message:
XHAUST returned error code -999 in: leaps.exhaustive(a, really.big really.big)
The library works just fine on some data (some work I did previously
that I ran again so that I could make sure that there wasn't some sort
of file problem), but there appears to be something wrong with this
particular data set. However, having examined the dataset, there are no
missing data and all of the data is numeric (shown below).
I looked in some of the archives, the online help, and some of the other
documentation, already.
Any ideas or suggestions would be very appreciated,
Thanks in advance,
Pete
-----------------------------------------------
> objects()
[1] "lahodsub" "last.warning" "pred"
"resp" > str(pred)
`data.frame': 93 obs. of 20 variables:
$ LDOC : num 0.613 0.672 0.556 0.906 0.958 ...
$ LTP : num 0.751 0.924 0.791 1.078 1.139 ...
$ LChlA : num 0.232 0.211 0.084 0.611 0.321 ...
$ LZmax : num 1.20 1.20 1.20 1.30 1.30 ...
$ LDepth : num 0.954 1.000 0.845 0.778 0.845 ...
$ LDSAm : num 6.13 6.13 6.13 6.79 6.79 ...
$ LLSAm : num 5.58 5.58 5.58 5.79 5.79 ...
$ LCALA : num 0.553 0.553 0.553 0.999 0.999 ...
$ LQS1 : num 0.400 0.400 0.400 0.781 0.781 ...
$ LRETENT1 : num 0.342 0.342 0.342 0.405 0.405 ...
$ LHAvT : num 0.941 0.865 1.007 0.874 0.872 ...
$ LLakeVol : num 6.36 6.36 6.36 6.46 6.46 ...
$ LLakeVSA : num 0.778 0.778 0.778 0.668 0.668 ...
$ LHypoVol : num 5.49 5.32 5.76 5.90 5.79 ...
$ LHypoVSA : num 0.455 0.389 0.549 0.625 0.629 ...
$ LHypoThick: num 0.875 0.813 0.978 1.161 1.130 ...
$ LVHVT : num -0.872 -1.035 -0.600 -0.565 -0.666 ...
$ LWRT : num -0.293 -0.293 -0.293 -0.740 -0.740 ...
$ LVTCA : num 0.227 0.227 0.227 -0.331 -0.331 ...
$ LVHCA : num -0.644 -0.807 -0.372 -0.896 -0.997 ...
> str(resp)
`data.frame': 93 obs. of 4 variables:
$ LAHOD : num -0.869 -1.254 -1.220 -0.820 -1.161 ...
$ LAHOD4C: num -1.011 -1.354 -1.405 -0.925 -1.265 ...
$ LVOD : num -1.32 -1.64 -1.77 -1.45 -1.79 ...
$ LVOD4C : num -1.47 -1.74 -1.95 -1.55 -1.89 ...
> library(leaps)
> lvodsub <- leaps(pred, resp$LVOD)
Warning message:
XHAUST returned error code -999 in: leaps.exhaustive(a) > rm(lvodsub)
> lvodsub <- subsets(pred, resp$LVOD)
Warning message:
XHAUST returned error code -999 in: leaps.exhaustive(a, really.big really.big)
--
Pete St. Onge
pete at seul.org
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On Mon, 25 Oct 1999, Pete St. Onge wrote:> Hi there, > This problem has been dogging me for a bit, and I'm trying to > figure out why. When running the the subsets function in the leaps > library, R is giving me the following error message > > > lvodsub <- subsets(pred, resp$LVOD) > Warning message: > XHAUST returned error code -999 in: leaps.exhaustive(a, really.big > really.big)The really.big option isn't relevant -- R is just reporting where the error was. Looking at the code it seems that error -999 occurs when the the QR decomposition that subsets constructs is (nearly) not of full rank. This shouldn't happen, since redundant variables are supposed to be eliminated earlier in the process. If you could send me the data I could try to find the problem. -thomas Thomas Lumley Assistant Professor, Biostatistics University of Washington, Seattle> > but this still happens if I add the really.big option: > > > lvodsub <- subsets(pred, resp$LVOD, really.big=T) > Warning message: > XHAUST returned error code -999 in: leaps.exhaustive(a, really.big > really.big) > > The library works just fine on some data (some work I did previously > that I ran again so that I could make sure that there wasn't some sort > of file problem), but there appears to be something wrong with this > particular data set. However, having examined the dataset, there are no > missing data and all of the data is numeric (shown below). > > I looked in some of the archives, the online help, and some of the other > documentation, already. > > Any ideas or suggestions would be very appreciated, > > Thanks in advance, > > Pete > > ----------------------------------------------- > > > objects() > [1] "lahodsub" "last.warning" "pred" "resp" > > str(pred) > `data.frame': 93 obs. of 20 variables: > $ LDOC : num 0.613 0.672 0.556 0.906 0.958 ... > $ LTP : num 0.751 0.924 0.791 1.078 1.139 ... > $ LChlA : num 0.232 0.211 0.084 0.611 0.321 ... > $ LZmax : num 1.20 1.20 1.20 1.30 1.30 ... > $ LDepth : num 0.954 1.000 0.845 0.778 0.845 ... > $ LDSAm : num 6.13 6.13 6.13 6.79 6.79 ... > $ LLSAm : num 5.58 5.58 5.58 5.79 5.79 ... > $ LCALA : num 0.553 0.553 0.553 0.999 0.999 ... > $ LQS1 : num 0.400 0.400 0.400 0.781 0.781 ... > $ LRETENT1 : num 0.342 0.342 0.342 0.405 0.405 ... > $ LHAvT : num 0.941 0.865 1.007 0.874 0.872 ... > $ LLakeVol : num 6.36 6.36 6.36 6.46 6.46 ... > $ LLakeVSA : num 0.778 0.778 0.778 0.668 0.668 ... > $ LHypoVol : num 5.49 5.32 5.76 5.90 5.79 ... > $ LHypoVSA : num 0.455 0.389 0.549 0.625 0.629 ... > $ LHypoThick: num 0.875 0.813 0.978 1.161 1.130 ... > $ LVHVT : num -0.872 -1.035 -0.600 -0.565 -0.666 ... > $ LWRT : num -0.293 -0.293 -0.293 -0.740 -0.740 ... > $ LVTCA : num 0.227 0.227 0.227 -0.331 -0.331 ... > $ LVHCA : num -0.644 -0.807 -0.372 -0.896 -0.997 ... > > > str(resp) > `data.frame': 93 obs. of 4 variables: > $ LAHOD : num -0.869 -1.254 -1.220 -0.820 -1.161 ... > $ LAHOD4C: num -1.011 -1.354 -1.405 -0.925 -1.265 ... > $ LVOD : num -1.32 -1.64 -1.77 -1.45 -1.79 ... > $ LVOD4C : num -1.47 -1.74 -1.95 -1.55 -1.89 ... > > > library(leaps) > > lvodsub <- leaps(pred, resp$LVOD) > Warning message: > XHAUST returned error code -999 in: leaps.exhaustive(a) > > rm(lvodsub) > > lvodsub <- subsets(pred, resp$LVOD) > Warning message: > XHAUST returned error code -999 in: leaps.exhaustive(a, really.big > really.big) > > > -- > Pete St. Onge > pete at seul.org > -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- > r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html > Send "info", "help", or "[un]subscribe" > (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch > _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._ >Thomas Lumley ------------------------------------------------------+------ Biostatistics : "Never attribute to malice what : Uni of Washington : can be adequately explained by : Box 357232 : incompetence" - Hanlon's Razor : Seattle WA 98195-7232 : : ------------------------------------------------------------ -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On Mon, 25 Oct 1999, Pete St. Onge wrote:> Hi there, > This problem has been dogging me for a bit, and I'm trying to > figure out why. When running the the subsets function in the leaps > library, R is giving me the following error message > > > lvodsub <- subsets(pred, resp$LVOD) > Warning message: > XHAUST returned error code -999 in: leaps.exhaustive(a, really.big > really.big)This turned out to be a design matrix with five out of twenty very nearly redundant columns. subsets() tries to predict whether it will run across a singularity at some point, but its prediction wasn't sufficiently paranoid. The problem should be very rare, but I will soon be releasing a version of the package that does glms (approximately) and I'll adjust the singularity tolerance then. Thomas Lumley Assistant Professor, Biostatistics University of Washington, Seattle -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._