There's now a new version of randomForest (3.3-4) on CRAN. Changes from the last version include: o "Unsupervised learning" now works again. o Fixed a bug in the original Fortran: the mtry variables chosen to split each node are now sampled without replacement. o Variable importance for regression now works correctly (i.e., reproducible when random seed is set). The problem was simply un-initialized array. The problem is also in the original Fortran code, so users of that beware! o A new option keep.forest (default TRUE), which, when set to FALSE, supresses the `forest' component of the returned object. This can reduce memory usage and execution time significantly, if prediction of test set is not needed. o The default value of mtry is now set to sqrt(p) for classification and p/3 for regression (where p is the number of predictors), consistent with Prof. Breiman's suggestion. Questions/comments/bug-reports/patches welcomed! Regards, Andy Andy I. Liaw, PhD Biometrics Research Phone: (732) 594-0820 Merck & Co., Inc. Fax: (732) 594-1565 P.O. Box 2000, RY84-16 Rahway, NJ 07065 mailto:andy_liaw at merck.com ------------------------------------------------------------------------------ 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 in 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. ============================================================================= -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-announce 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-announce-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._