Hi Andr?
>> I'm trying to use bart() ... but I can't predict my test data.
There is no predict() method, as such, but read ?pdbart very carefully, and
work through the examples. The fitted/predicted values from a fit are in
bartobject$yhat.train.mean
It would be a good idea to read the two PDFs that the authors provide at
http://faculty.chicagogsb.edu/robert.mcculloch/research/rob-mcculloch-cv.html
I've had excellent results from BayesTree, with fits that rival those of
Friedman's RuleFit, though the whole process takes a great deal longer.
HTH,
Mark.
andreBR wrote:>
> Hi all,
>
> sorry for my english, but I don't speak yours language.
> I'm trying to use bart() and rbf(). The package I'm using now is
> "BayesTree" and "neural", respectively. I could create
the models, but I
> can't predict my test data.
> Does anyone have such an experience? Any advice is appreciated!
>
> Thank you in advanced!.
>
>
> Andr?
>
--
View this message in context:
http://www.nabble.com/How-to-use-BayesTree-or-RBF-for-predict-tp15562371p15562524.html
Sent from the R help mailing list archive at Nabble.com.