Some R functions do not work for a MaxEntReplicates object

[Daniele Da Re]

Dear all, 

I am running some maxent models into R using package "dismo". 

mx1 <- maxent(bio.cr_main, train_p, train_a, args=c("-J", "-P", "maximumiterations=1000", "-A"))

mx2 <- maxent(bio.cr_main, train_p, train_a, args=c("-P","replicates=5","maximumiterations=1000","-A"))

I have found that for object mx1 the functions plot(), evaluate() and predict() work without problems, but they return an error message with a MaxEntReplicates object like mx2. 

I think because using "replicates" arguments the functions above do not know which of the 5 object use.

Is there a way to specify one of the replicates, in particular the mean values and standard deviations?

thank you for any advice,

Daniele 

[Jamie M. Kass]

Just a question, but why are you using replicates in the first place? Are you trying different random background samples each time? If you have a good background sample, in that there is little variation across samples, each replicate should be very similar.

Jamie Kass
PhD Candidate
City College of NYC

[Daniele Da Re]

Thank you for your reply Jamie!
Well, I don't have a good background sample, I have random background points. I was using "replicates argument" to get a Standard deviation map of the prediction.

[ndimhypervol]

 The random sample should be a good representation of the background, and if it's not, you can increase the number of background points you sample. You should also make sure you are specifying the background region appropriately -- if it's too large, you will likely need a very large background sample, but more importantly, if you include areas the species is unable to disperse to because of biogeographical barriers, etc., you are sending the model a false signal, i.e. these areas will be considered unoccupied and therefore low suitability even though the conditions are actually good for the species. Thus, think about restricting your background region, and inflating the number of points, and you should have a better overall sample.

Jamie Kass

PhD Candidate, CCNY

cell: 917-602-5787

--
You received this message because you are subscribed to a topic in the Google Groups "Maxent" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/maxent/jCg3G5sBdt4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to maxent+un...@googlegroups.com.
To post to this group, send email to max...@googlegroups.com.
Visit this group at https://groups.google.com/group/maxent.
For more options, visit https://groups.google.com/d/optout.