Thanks Torsten, That seems to fit the bill nicely!
Best, Florian Florian Koller-Meinfelder Research Consulting & Development ______________________________ GfK Fernsehforschung GmbH Nordwestring 101 90319 Nürnberg Tel +49 (0)911 395-3554 Fax +49 (0)911 395-4130 www.gfk.com/gfkfernsehforschung Torsten Hothorn <[EMAIL PROTECTED]> 19.04.2007 14:46 Bitte antworten an [EMAIL PROTECTED] An Florian Koller-Meinfelder <[EMAIL PROTECTED]> Kopie [email protected] Thema Re: [R] "tree-ID" in any segmentation package available? On Thu, 19 Apr 2007, Florian Koller-Meinfelder wrote: > Dear R-helpers, > > I am looking for a segmentation package that gives some "tree identifier" > as output for every observation in the data set (my response variable is > binary). I have skimmed through "rpart", "ada" and "adabag": The output > "trees" gives you the formula, but I have to run several thousand > segmentations on different data sets and it is tricky to use this > information within a macro (the only thing I could think of is to use some > string manipulation on the tree formula and apply it to the data, but I > hope there is an easier way - e.g. if the algorithm created 12 different > trees a vector that links every observation to one of these 12 segments > would be ideal). > is this > library("party") > airq <- subset(airquality, !is.na(Ozone)) > airct <- ctree(Ozone ~ ., data = airq, + controls = ctree_control(maxsurrogate = 3)) > where(airct) [1] 5 5 5 5 5 5 5 5 3 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 3 5 6 9 9 6 5 5 5 5 5 8 9 [38] 6 8 9 8 8 8 8 5 6 6 3 6 8 8 9 3 8 8 6 9 8 8 8 6 3 6 6 8 8 8 8 9 8 9 6 6 5 [75] 3 5 6 6 5 5 6 3 8 9 8 8 8 8 8 8 8 8 9 6 6 5 5 6 5 3 5 5 3 5 5 5 6 5 5 6 5 [112] 5 3 5 5 5 what you want? `where' gives you the number of the terminal node each observation in the learning sample is element of. Best wishes, Torsten > Cheers, > Florian > > > > > Florian Koller-Meinfelder > Research Consulting & Development > ______________________________ > > GfK Fernsehforschung GmbH > Nordwestring 101 > 90319 Nürnberg > > Tel +49 (0)911 395-3554 > Fax +49 (0)911 395-4130 > www.gfk.com/gfkfernsehforschung > > > > > > This email and any attachments may contain confidential or...{{dropped}}
______________________________________________ [email protected] 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.
