[R-sig-Geo] Dismo Maxent - Error with evaluate(), missing layers (or wrong names)

[Johannes Radinger]

Hi,

I am running multiple Maxent models for a list of 40 species in a loop.
Within the loop I am splitting the presence points into a train dataset
and a test dataset. All the model fitting and the prediction works fine,
there is just a problem with some species and evaluate().
What I am doing and the error:

#witholding a 20%(1/5) sample for testing 
> fold <- kfold(subset.occ, k=5)
> occtest <- subset.occ[fold == 1, ]
> occtrain <- subset.occ[fold != 1, ]

> occtest
            coordinates              species
4025 (3525400, 6055500) N.Gesamt.Meerforelle
> occtrain
            coordinates              species
4021 (3527250, 6060480) N.Gesamt.Meerforelle
4023 (3526100, 6058550) N.Gesamt.Meerforelle
4056 (3522320, 6048620) N.Gesamt.Meerforelle
4066 (3526840, 6060240) N.Gesamt.Meerforelle
4090 (3527920, 6063450) N.Gesamt.Meerforelle

> evaluate(me, p=occtest, a=bg, x=predictors)
Fehler in .local(object, ...) : missing layers (or wrong names)


I guess that is related to the number of occtest-samples. As there
is only one sample I might get this error. I just wanted
to get a proof if I am right with this guess, as the error message
is difficult to interpret to me.

thank you!

cheers,
/johannes



-- 

Jetzt informieren: http://mobile.1und1.de/?ac=OM.PW.PW003K20328T7073a