[OSGeo-Discuss] Random Forest Classification
dan.putler at sauder.ubc.ca
Wed Oct 26 00:06:01 PDT 2011
This seems to be an R topic rather than an OSGeo topic. The R-Geo
mailing list is probably the best choice (look at the mailing list page
on the http://www.r-project.org site). Random forest models can be used
with data acquired by remote sensing, but the method is really aspatial
in nature (the ability to deal with any autocorrelation, including
spatial autocorrelation, is pretty limited in decision tree based
methods, and random trees are an ensemble of decision trees, which
create a "forest"). At this point, I'd do some more homework on decision
trees (also commonly called classification and regression trees, or
CART, models) so you have a better sense of the underlying concepts, and
then hit the R-Geo mailing list if you have additional questions.
On 10/25/2011 11:40 PM, Mohammed Rashad wrote:
> Hi All,
> I wrant to do Random Forest classification. I installed R,
> randomForest classifier package for R
> but dont know how to use it.
> Is there any Open Source Remote sensing application which do RF
> classification on satellite images?
> Anyone r has random forest classification example?
> Any language or package example no problem.
> Does anyone did it in R?
> if yes how?
> I google RF Classification but most of them are for medical disease
> and research not for Remote Sensing
> Mohammed Rashad K M
> M.S. (By Research) student
> Lab for Spatial Informatics
> Department of CSE
> International Institute of Information Technology
> Hyderabad, India
More information about the Discuss