[Biomod-commits] Warning messages from BIOMOD_Modeling function

[Damien Georges]

Dear Marco,

When I tried to execute the script you posted, I only get the 2 warnings 
referring to GLM issues (5th and 6th of your warnings).

That's quite common warnings that append when glms not managed to fully 
converged. I guess it's not so alarming in our case.

This warnings messages concerned GLM algorithms because you have set the 
option "rescal.all.model" to TRUE. This option implies all models built 
to be rescaled by a binomial glm. That ensures all models prediction to 
be on a same scale and comparable one to another.
Setting this option to FALSE will remove this warnings.

All the best,

Damien.


On 15/10/2012 19:20, Marco Girardello wrote:
> Dear BIOMOD list,
>
> I am trying use a couple of methods from the BIOMOD_Modeling function
> (biomod2), but I get some warning messages which I do not understand.
> Below you will find a reproducible example. The surprising thing about
> the warnings is that at least two of them are referred to the method
> "GLM", which I did not use for the analysis.  Is this a bug? Or
> something that I do not understand?
>
> any help would be greatly appreciated!
>
>
> Best wishes,
>
> Marco
>
>
>
>
> ####
> dat<-read.csv("http://dl.dropbox.com/u/33966347/spec2.csv")
>
> myBiomodData <- BIOMOD_FormatingData(resp.var = as.numeric(dat$resp),
>                                         expl.var = dat[4:19],
>                                         resp.xy = dat[1:2],
>                                         resp.name = "spec2")
>
>
> myBiomodOption <-
> BIOMOD_ModelingOptions(GLM=NULL,GBM=list(n.trees=2000),GAM=NULL,CTA=NULL,ANN=NULL,SRE=NULL,FDA=NULL,MARS=NULL,MAXENT=NULL)
>
>
> ### Modelling
> myBiomodModelOut <- BIOMOD_Modeling(
>                               myBiomodData,
>                               models = c("GBM","RF"),
>                               models.options = myBiomodOption,
>                               NbRunEval=1,
>                               DataSplit=80,
>                               Yweights=NULL,
>                               VarImport=0,
>                               models.eval.meth = c('TSS','ROC'),
>                               SaveObj = TRUE,
>                               rescal.all.models = TRUE)
>
>
> Warning messages:
> 1: In max(i) : no non-missing arguments to max; returning -Inf
> 2: In max(i) : no non-missing arguments to max; returning -Inf
> 3: In max(i) : no non-missing arguments to max; returning -Inf
> 4: In max(i) : no non-missing arguments to max; returning -Inf
> 5: glm.fit: algorithm did not converge
> 6: glm.fit: fitted probabilities numerically 0 or 1 occurred
> 7: In is.na(object) :
>    is.na() applied to non-(list or vector) of type 'NULL'
> _______________________________________________
> Biomod-commits mailing list
> Biomod-commits at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/biomod-commits