Errors in GLM, BRT and RF - datasets too small?

[virginia...@gmail.com]

Hello!

I've been getting the following errors for two species occurrence datasets and was wondering if there is anything I can do to address them/get the modules to complete successfully or if the errors are likely the result of having too small species occurrence datasets/a result of something related to data structure. At this point I have just removed these modules from the workflows for these species and was going to work with the output from the other modules:

GLM module - dataset is presence/absence (71 presence; 62 absence):

An error was encountered in the R script for this module.
     The R error message is below:
Error in generic.model.fit(out, Model, t0) : Null model was selected. 

Evaluation metrics and plots will not be produced

Calls: FitModels -> generic.model.fit

Execution halted

BRT module - dataset is presence only (7 presence). I was able to get rid of this error by decreasing the background sample size from 10,000 to a number closer to the number of presence points for two other datasets with larger numbers of presences (29 and 30). My assumption is that this dataset is just too small and unless I can get more presence points, the BRT and RF algorithms won't be able to complete:

 

An error was encountered in the R script for this module.
     The R error message is below:
Error in gbm.fit(x, y, offset = offset, distribution = distribution, w = w,  :

  The dataset size is too small or subsampling rate is too large: nTrain*bag.fraction <= n.minobsinnode

Calls: FitModels ... est.lr -> gbm.step.fast -> eval -> eval -> gbm -> gbm.fit

Execution halted

RF module - dataset is presence only (7 presence):

An error was encountered in the R script for this module.
     The R error message is below:
Error in if (Improve > improve) { : missing value where TRUE/FALSE needed

Calls: FitModels -> generic.model.fit -> model.fit -> tuneRF

Execution halted

Any thoughts are much appreciated.

Virginia