[R-sig-eco] Slow modelisation in Biomod2 package

Rajendra Mohan Panda rmp.iit.kgp at gmail.com
Mon Jul 11 08:33:12 CEST 2016


Dear All

I am using on BIOMOD2 package for species distribution emsemble. I have
taken five models (GLM, CTA, RF, GBM, ANN), and fitted 12 environmental
variables.

My data points extend upto 9076.

I have increased memory size to 2g using 'rjava'
options(java.parameters = "-mex2g")



I have used following codes to run my model:

myBiomodModelOut <- BIOMOD_Modeling (myBiomodData,
                                     models = c('GLM', 'GBM', 'CTA', 'ANN',
'RF'),
                                     models.options = myBiomodOption,
NbRunEval=1,
                                     DataSplit=80, Prevalence=0.5,
                                     VarImport=3, models.eval.meth =
c('TSS', 'ROC'),
                                     SaveObj = TRUE, rescal.all.models =
TRUE,
                                     do.full.models = FALSE, modeling.id =
"test")

However, model response is very slow. Modelisation does not complete even
after three days (>72 hrs).
Can anybody explain the way to accelerate the process?

I will be thankful for your valuable advice.

Best Regards
Rajendra M Panda
School of Water Resources
Indian Institute of Technology Kharagpur

	[[alternative HTML version deleted]]



More information about the R-sig-ecology mailing list