[R-sig-Geo] Using cokriging to predict the probability of a binary variable

Joel Thomas joeliol at gmail.com
Mon Dec 14 22:04:14 CET 2009


Hi,
I am trying to create a map showing the probability of the presence of
a binary variable. All of the data points I have are known locations
of what I'm trying to find with several topographic variables as
predictors. I could create a set of random points and populate their
attributes with the same predictors, but the problem is I can't assign
a probability of there being what I'm trying to find - its either Yes
or No. Is it possible to interpolate between points using cokriging on
some kind of false scale when the points I'm trying to interpolate
between all have the same target value, ie. Yes or 100% probability?

Thanks,
Joel



More information about the R-sig-Geo mailing list