randomForest crash?
Karen, There were a couple of stupid errors in 3.4-4. I've uploaded a patched release, 3.4-5, to CRAN. Your example now runs without problem. However, you probably are not using the function correctly, even though the new version runs without problem. The response you are using, V217, is numeric with values 1 and 2. randomForest() will think that you actually have numerical response, and thus fit a regression model, rather than classification. To do classification, the response *must* be factor. Last but not least, it's a good idea to upgrade R: if not for the new features, at least for the bug fixes. Cheers, Andy
From: Karen.Chancellor at asu.edu [mailto:Karen.Chancellor at asu.edu]
Hi Andy,
Here is the data.frame, and my command is
lung.rf <- randomForest(V217 ~ ., data=tlAS216cda,
importance=TRUE,
proximity=TRUE)
Thanks for checking it on your machine.
Also, I have R-1.5 on a linux6.1 machine. Will this version
of randomForests
run on that machine? If not, is there one that will?
Thanks.
Karen
Quoting "Liaw, Andy" <andy_liaw at merck.com>:
Hi Karen, I have not seen such problem, nor heard similar report.
All I can say
is that we have used the package on much larger data sets without problem. As the R developers say, the best way to get bugs fixed is
by crashing
the developer's machine. If you can send me the data (without
column/row
labels, if need be), I can try and see if it crashes R on
my machine.
Other than that, there's not much I can offer. Cheers, Andy
-----Original Message----- From: Karen.Chancellor at asu.edu [mailto:Karen.Chancellor at asu.edu] Sent: Monday, April 21, 2003 2:57 AM To: r-help at stat.math.ethz.ch Subject: [R] randomForest crash? I am attempting to use randomForests to look for interesting genes
in
microarray data with 216genes, 2 classes and 52 samples. My
data.frame
is 52x217 with the last column, V217 being the class(1 or 2).
When I try
lung.rf <- randomForest(V217 ~ ., data=tlSA216cda,
importance=
TRUE, proximity = TRUE) the GUI crashes. I am running R-1.6.2 under windo$e98, and most recent version of randomForest. Your ideas appreciated. Karen .- --. ....- -.-. -.-.
______________________________________________ R-help at stat.math.ethz.ch mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
-------------------------------------------------------------- ----------------
Notice: This e-mail message, together with any attachments, contains information of Merck & Co., Inc. (Whitehouse Station, New
Jersey, USA)
that may be confidential, proprietary copyrighted and/or legally privileged, and is intended solely for the use of the individual or entity named on this message. If you are not the intended
recipient,
and have received this message in error, please immediately
return this
by e-mail and then delete it.
============================================================== ================