R help - Jul 2010 - Is there a "Mixed effect model" for regression/classification trees?

Hello all,

This is more a statistical question then an R question, but I am sure it
will have an R interpretation to it.

If I wish to predict an outcome based on some potential features, I could
(in some cases) use either regression or regression-tree.
However, if my observations are divided to groups (for example by
"subject"), I might then want to model that using a random effect for
the
"subject" and a fixed effect for the other features I wish to model
for the
prediction.
My question is what (if exist) is the parallel of this in regression trees ?
Is it simply like adding the "subject" classifier to the tree? or is
this
leading to a different model scheme all together? (and if so - what is it)


Thanks,
Tal




----------------Contact
Details:-------------------------------------------------------
Contact me: Tal.Galili@gmail.com |  972-52-7275845
Read me: www.talgalili.com (Hebrew) | www.biostatistics.co.il (Hebrew) |
www.r-statistics.com (English)
----------------------------------------------------------------------------------------------

	[[alternative HTML version deleted]]

On Sun, 11 Jul 2010, Tal Galili wrote:
> Hello all,
>
> This is more a statistical question then an R question, but I am sure it
> will have an R interpretation to it.
>
> If I wish to predict an outcome based on some potential features, I could
> (in some cases) use either regression or regression-tree.
> However, if my observations are divided to groups (for example by
> "subject"), I might then want to model that using a random effect
for the
> "subject" and a fixed effect for the other features I wish to
model for the
> prediction.
> My question is what (if exist) is the parallel of this in regression trees
?
> Is it simply like adding the "subject" classifier to the tree? or
is this
> leading to a different model scheme all together? (and if so - what is it)
We had thought about doing something in this direction in the model-based 
recursive partitioning framework (MOB) in the "party" package but
never
pulled it off.

However, I've seen various other ideas at conferences for trees with 
random intercepts etc.

One CRAN package that provides infracstructure for this is "REEMtree" 
which is probably worth a look.

hth,
Z
>
> Thanks,
> Tal
>
>
>
>
> ----------------Contact
> Details:-------------------------------------------------------
> Contact me: Tal.Galili at gmail.com |  972-52-7275845
> Read me: www.talgalili.com (Hebrew) | www.biostatistics.co.il (Hebrew) |
> www.r-statistics.com (English)
>
----------------------------------------------------------------------------------------------
>
> 	[[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.
>