[R] randomForest interaction detection

From: Retzer Joe <retzerjj_at_gmail.com>
Date: Wed, 30 Apr 2008 11:04:01 -0500


I've been using the randomForest package and getting very useful results for some time now.

I have a question pertaining to variable interactions. Since, unlike CART, no single tree representative of the solution is available, I was wondering if there was some way of discerning / identifying potential interactions in the data. I'll sometimes produce 2-way partial dependence plots but I select variables for these plots somewhat arbitrarily (i.e. without prior suggestion of interaction).

On Breimansí website he mentions an experimental approach to identify interactions. I know this has come up before, a couple years ago, and Andy Liaw noted it was not implemented in part due to its experimental nature.

I was wondering if anyone had a suggestion for detecting interactions using the randomForest package (In an analogous manner to Breimans' suggestion or another approach).

Many Thanks,
Joe Retzer

        [[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. Received on Wed 30 Apr 2008 - 17:29:25 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Wed 30 Apr 2008 - 17:30:34 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive