Do any of you know if there is an R implementation of logistic regression trees?
Specifically, I am looking for a logistic regression tree implementation in R based on a specific algorithm (LOTUS - Logistic regression Tree with Unbiased Splits) developed by Kin-Yee Chan and Wei-Yin Loh.
P.S: I am familiar with LMT (Logistic Model Trees) available through the RWeka package and mob (party package).
I wouldn't know of an R implementation of LOTUS.
The binaries for the original LOTUS implementation are available from Kin-Yee Chan's web page (http://www.stat.nus.edu.sg/~kinyee/lotus.html) but I think Wei-Yin Loh's recommendation nowadays is to use GUIDE instead (http://www.stat.wisc.edu/~loh/guide.html). A while back, when developing the MOB algorithm, we had written an R interface to the GUIDE binaries. However, tracking the control options from version to version turned out to be quite cumbersome necessitating too much reverse engineering. So we gave it up again.
As for MOBs based on logistic regression: It is much simpler to use glmtree(..., family = binomial) from partykit rather than the original mob() implementation in party.
