[R-sig-Geo] Sampling polygons for analysis in CART/randomForests

[Tim Howard]

The lack of response indicates I was straying too far from the intent of
the list. I apologize for my momentary lack of judgement... it won't
happen again.

Sincerely,
Tim Howard

****
Dear R - Geo community:
   This is perhaps a bit off topic (really a geostats question), but
we've been struggling with these questions for about 6 months now, and
it just reared its head again when I posed a question on one of ESRI's
discussion groups. I realized you would be much more experienced in
this
issue and might be interested in offering advice. And this might be
something others are struggling with as well. 
   Background: We are creating predictive habitat distribution models
(a good synthesis paper is Guisan and Zimmerman 2000. Ecological
Modelling 135:147-186); using known locations of species to create a

<snip>