[Biomod-commits] "xzfile" error with models function

[Wilfried Thuiller]

Dear Sarah,
Sorry it took me so long to answer. I have no clue about the problem but I expect it has something to do with the selection of pseudo-absence.
Please have a try with using "random" instead and let me know.

In the meantime, you seem to have a very very large number of variables. Given the names, I expect that many of them are strongly correlated. Running most of these models on correlated data is not a very good practice. 

Cheers,
Wilfried


Le 22 janv. 2011 à 03:34, Sarah Skikne a écrit :

> Dear BIOMODers,
> 
> I am having some trouble with the models() function in BIOMOD and was hoping to get some help.  
> 
> Here is the relevant info:
> - R version 2.12.1
> - I re-downloaded the BIOMOD package a few days ago so it should be up to date.
> - code:
> 
> TrialDF<-data.frame(XCoord=x, YCoord=y, RAIN1=RAIN1scan, RAIN2=RAIN2scan, RAIN3=RAIN3scan, RAIN4=RAIN4scan, RAIN5=RAIN5scan, RAIN6=RAIN6scan, RAIN7=RAIN7scan, RAIN8=RAIN8scan, RAIN9=RAIN9scan, RAIN10=RAIN10scan, RAIN11=RAIN11scan, RAIN12=RAIN12scan, TMAX1=TMAX1scan, TMAX2=TMAX2scan, TMAX3=TMAX3scan, TMAX4=TMAX4scan, TMAX5=TMAX5scan, TMAX6=TMAX6scan, TMAX7=TMAX7scan, TMAX8=TMAX8scan, TMAX9=TMAX9scan, TMAX10=TMAX10scan, TMAX11=TMAX11scan, TMAX12=TMAX12scan, TMEAN1=TMEAN1scan, TMEAN2=TMEAN2scan, TMEAN3=TMEAN3scan, TMEAN4=TMEAN4scan, TMEAN5=TMEAN5scan, TMEAN6=TMEAN6scan, TMEAN7=TMEAN7scan, TMEAN8=TMEAN8scan, TMEAN9=TMEAN9scan, TMEAN10=TMEAN10scan, TMEAN11=TMEAN11scan, TMEAN12=TMEAN12scan, TMIN1=TMIN1scan, TMIN2=TMIN2scan, TMIN3=TMIN3scan, TMIN4=TMIN4scan, TMIN5=TMIN5scan, TMIN6=TMIN6scan, TMIN7=TMIN7scan, TMIN8=TMIN8scan, TMIN9=TMIN9scan, TMIN10=TMIN10scan, TMIN11=TMIN11scan, TMIN12=TMIN12scan,  LOCALITY=localityscan)
> 
> Initial.State(Response=TrialDF[,51], Explanatory=TrialDF[,3:50], IndependentResponse=NULL, IndependentExplanatory=NULL, sp.name='spp1')
> 
> 
> 
> Models(GLM=T, TypeGLM="poly", Test="AIC", GBM=T, No.trees=2000, GAM=T, CTA=T, CV.tree=50, ANN=T, CV.ann=3, SRE=T, MARS=T, RF=T, NbRunEval=3, DataSplit=80, NbRepPA=2, strategy="circles", coor=TrialDF[,1:2], distance=.5, nb.absences=220, Yweights=NULL, VarImport=T, Roc=T, Optimized.Threshold.Roc=T, Kappa=T, TSS=T, KeepPredIndependent=T)
> 
> 
> From this, I get the following:
> 
> Modelling summary
> 
> -----------------------------------
> 
> Number of species modelled :            1
> 
> redwood
> 
> numerical variables :                   RAIN1, RAIN2, RAIN3, RAIN4, RAIN5, RAIN6, RAIN7, RAIN8, RAIN9, RAIN10, RAIN11, RAIN12, TMAX1, TMAX2, TMAX3, TMAX4, TMAX5, TMAX6, TMAX7, TMAX8, TMAX9, TMAX10, TMAX11, TMAX12, TMEAN1, TMEAN2, TMEAN3, TMEAN4, TMEAN5, TMEAN6, TMEAN7, TMEAN8, TMEAN9, TMEAN10, TMEAN11, TMEAN12, TMIN1, TMIN2, TMIN3, TMIN4, TMIN5, TMIN6, TMIN7, TMIN8, TMIN9, TMIN10, TMIN11, TMIN12
> 
>  
> number of evaluation repetitions :      3
> 
> number of pseudo-absences runs :        2
> 
> models selected :                       ANN, CTA, GAM, GBM, GLM, MARS, RF, SRE
> 
> total number of model runs :            64
> 
> -----------------------------------
> 
>  
> #####                    redwood                        #####
> 
> #####              pseudo-absence run 1                 #####
> 
> Model=Artificial Neural Network
> 
>          3 Fold Cross Validation + 3 Repetitions
> 
> Calibration and evaluation phase: Nb of cross-validations:  3
> 
> Error in xzfile(file, "wb", compression = 9) : cannot open the connection
> 
> In addition: There were 15 warnings (use warnings() to see them)
> 
> 
> 
> This code used to work for me so not sure what has changed.  When I try it using the sample data given in the manual, I get the error about 50% of the time.  When I use my data, I get the error every time.
> 
> 
> 
> Any help on the "xzfile" error and how to avoid it would be great.  
> 
> 
> 
> Thanks!
> 
> Sarah
> 
> 
> 
> 
> 
> Center for Applied Biodiversity Informatics
> 
> California Academy of Sciences
> 
> 
> 
> 
> _______________________________________________
> 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

--------------------------
Dr. Wilfried Thuiller
Laboratoire d'Ecologie Alpine, UMR CNRS 5553
Université Joseph Fourier
BP53, 38041 Grenoble cedex 9, France
tel: +33 (0)4 76 51 44 97
fax: +33 (0)4 76 51 42 79

Email: wilfried.thuiller at ujf-grenoble.fr
Home page: http://www.will.chez-alice.fr
Website: http://www-leca.ujf-grenoble.fr/equipes/tde.htm

FP6 European MACIS project: http://www.macis-project.net
FP6 European EcoChange project: http://www.ecochange-project.eu





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/biomod-commits/attachments/20110203/e887c977/attachment.htm>