[R] Adjusted R2 for Multivariate Regression Trees (MRT) (ignore the previous message)
David L Carlson
dcarlson at tamu.edu
Tue Sep 2 16:21:20 CEST 2014
You should probably pose this question directly to the package author as the function you are using cites an unpublished manuscript as a reference. It appears that the function uses random draws to estimate R2 so each result is approximately correct and no result is exactly correct. You can probably take the mean of the 100 runs as a reasonable estimate. If the estimates are quite variable, you should probably use more than 40 runs by setting T=100 or an even larger number. Multiple runs should then be more similar to one another.
-------------------------------------
David L Carlson
Department of Anthropology
Texas A&M University
College Station, TX 77840-4352
-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of Jackson Rodrigues
Sent: Monday, September 1, 2014 3:08 PM
To: r-help at r-project.org
Subject: [R] Adjusted R2 for Multivariate Regression Trees (MRT) (ignore the previous message)
Dear fellows,
I am using MVPARTwrap package to built a MRT of 25 pollen samples collected
from 5 different ecosystems, on my analysis I will include adjusted R2.
Based on MVPARTwrap package I want to get adjusted R2 for my MRT for this,
I am using the code below.
#step 1 - Building MRT.
Pre_euro.mvpart <- mvpart(data.matrix(mydata.2) ~ .,5Ecosystems,
margin=0.02, cp=0, xv="pick", xval=nrow(mydata.1), xvmult=100, which=4)
MRT.mite.tree<-MRT(Pre_euro.mvpart, 10, LABELS=LABELS)
#step 2- Adjusted R2
R2aGDF(MRT.mite.tree, T=40, tau_const=0.6, 5Ecosystems)
However, if run adjusted R2 code (step 2) 100 times, I will get 100
different results. Which one is correct?
Does anyone can help me? Any help is very welcome.
Cheers.
Jackson Rodrigues
[[alternative HTML version deleted]]
______________________________________________
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