[Biomod-commits] Please help; new with BIOMOD!

Andreas Soteriades andreassot10 at yahoo.com
Sat Feb 25 22:14:38 CET 2012


Hi all,

I am trying to run the MaxEnt tutorial datasets (presences + backgroung points with localities, bioclimatic variables and one categorical variable for Bradypus Variegatus) in BIOMOD and am facing some trouble:

1)

Models(GLM = T, TypeGLM = "poly", Test = "AIC",
+ GBM = T, No.trees = 2000,
+ GAM = T,Spline = 3, CTA = T, CV.tree = 50,
+ ANN = T, CV.ann = 2,
+ SRE = T, quant=0.025,
+ FDA = T,
+ MARS = T,
+ RF = T,
+ NbRunEval = 3, DataSplit = 80, Yweights=NULL,
+ Roc = T, Optimized.Threshold.Roc = T, Kappa = T, TSS=T,
+ KeepPredIndependent = T, VarImport=5,
+ NbRepPA=2, strategy="circles", coor=LatLong,
+ distance=2, nb.absences=1000)

Even though BIOMOD warns me about MARS' and SRE's inability to handle categorical variables:

MARS model was shut down, it cannot run on factorial variables : ecoreg
SRE model was shut down, it cannot run on factorial variables : ecoreg

I do not get any warnings for all the rest (ANN, CTA, GAM, GBM, GLM, FDA, RF). Nevertheless, the “ models” folder only contains results for CTA and ANN.

2)
No Biomod.PA.sample$Sp1$PA1 object is generated after running the models.

3)
I have no predictions in my pred folder.

4)
I want to evaluate the importance of each variable according to the procedure described in BIOMOD tutorial, 2012 p.30:
 
model <- glm(Sp281 ~ Var1 + Var2 + Var3 + Var4 + Var5 + Var6 + Var7, data=Sp.Env)
Pred <- predict(model, Expl.Var, type="response")
Expl.Var2 <- Expl.Var
Expl.Var2[,'Var1'] <- sample(Expl.Var[,'Var1'])
Pred2 <- predict(model, Expl.Var2, type="response")
par(mfrow=c(1,2))
level.plot(Pred, LatLong, show.scale=F, cex=0.8)
level.plot(Pred2, LatLong, show.scale=F, cex=0.8)
 
First of all, does someone have to go through the whole procedure of typing the code above for each model and for each variable, and to use the predict() function each time? According to the tutorial “once the models are trained (i.e. calibrated), a standard prediction is made. Then, one of the variables is randomized and a new prediction is made.” I thought that BIOMOD was following this procedure automatically. If yes, how? If not, how can I load CTA, ANN (and all the rest of the models in general) the same way glm is loaded in the code above (i.e. glm(Sp281 ~ Var1 + Var2 + Var3 + Var4 + Var5 + Var6 + Var7, data=Sp.Env))?

Cheers,

Andreas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/biomod-commits/attachments/20120225/993548dd/attachment.html>


More information about the Biomod-commits mailing list