Hello R-list!
I'm developing a prognostic model for a medical problem (event yes/no).
So I used the patrty package and did a Random Forest:
ctrl <- cforest_unbiased(ntree=1000, mtry=3, minsplit=100)
set.seed(675)
rf <- cforest( ph ~. , data=train, controls=ctrl)
print(rf)
and calculated the individual remissionprobabilities:
wktrf <- treeresponse(rf)
Now my questions:
1. What does treeresponse exactly do here?
Does it calculate the arithmetic mean of the class probabilities of the
end nodes the patient belongs to for all trees in the forest for each
patient?
2. As I want only two risk groups I used ctree on the probability for
the event which I got from treeresponse to get the optimal cutpoint. Is
this a good approach?
Maybe someone can help?
Thanks a lot,
Verena Hoffmann
______________________________________________
R-help@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.