[Biomod-commits] MaxEnt and BIOMOD

Tue Apr 26 06:50:30 CEST 2011

The main problem with Maxent (Maximum Entropy) is that this is a technique (like a GLM or Bayesian framework) and not the name of a model.

The Maxent you are refering to (in FD) is not made to a model species distributions but to retrieve abundance distribution within communities based on functional traits. 

Shipley, B., D. Vile, et al. (2006). From plant traits to plant communities: a statistical mechanistic approach to biodiversity. Science 314, 812-814.

Cheers,
Wilfried

Le 25 avr. 2011 à 23:39, Charles Novaes de Santana a écrit :

> Thank you very much for your answer, Wilfried!! That are really good
> news!! It is good to know that you are working on it!
> 
> Just a comment more: Do you know that there is a package in R (FD)
> that already have an implementation of MaxEnt? Would it be useful to
> implement MaxEnt for BIOMOD?
> 
> - http://cran.r-project.org/web/packages/FD/
> - http://cran.r-project.org/web/packages/FD/FD.pdf
> 
> Anyway, good luck with your great work!! I hope I can help you with
> this someday in the future (I am still starting to know the world of
> niche modeling... :) )
> 
> Charles
> 
> On Sat, Apr 23, 2011 at 10:38 AM, Wilfried Thuiller
> <wilfried.thuiller at ujf-grenoble.fr> wrote:
>> Dear Charles,
>> 
>> First of all, I must say that it is a great pleasure to read all the
>> interesting discussions we have in this list, we learn a lot with
>> every message! Thanks, and congratulations!!
>> Well, sorry if the question I will point here has been done before,
>> but I really didn't find anything about it in previous messages I
>> have:
>> 
>> Is it posible to include MaxEnt among the models simulated by BIOMOD?
>> 
>> In my opinion, the great strength of BIOMOD, beyond many other
>> benefits, is to allow the ensemble of many models in an easy way, and
>> so to give us more confidence results of our simulations. But I see
>> that MaxEnt is often used to simulate distribution of species and
>> sometimes I ask myself why it is not included in BIOMOD (yet). Is it a
>> "stupid" question? (I really believe that there are NO "stupid"
>> questions, but... :-))
>> 
>> This is indeed a very interesting question.
>> The answer is no and yes.
>> No because Maxent is coded in Java, much more efficient than in R. We'll
>> have to recode the entire program in C or C++ to fully integrate it in R and
>> latter in BIOMOD.
>> However, this is easy to run Maxent using a batch command from R. I have
>> thought about putting Maxent within BIOMOD that way since a while. I am just
>> to busy to do it but hope to start looking at that this year.
>> For curious, find below a short code for running Maxent from R.
>> You need the ascii data for running it and modify all paths and data for
>> your own data.
>> My species is Sp290.csv (Sp290 from SpEnv data in BIOMOD exported in csv
>> format for Maxent).
>> The environmental variables are the same than SpEnv, as many ascii files as
>> variables.
>> You need maxent.jar in your working directory.
>> dir.create(path=paste(getwd(), "/MaxEnt.res", sep=""))
>> MaxEnt.layers <- paste(getwd(), "/europe/ascii/current", sep="")
>> MaxEnt.out <- paste(getwd(), "/MaxEnt.res", sep="")
>> MaxEnt.samples <- paste(getwd(), "/europe/point", sep="")
>> layers <- list.files(MaxEnt.layers)
>> # run MaxEnt:
>> system(command=paste("java -jar maxent.jar environmentallayers=",
>> MaxEnt.layers, " samplesfile=", MaxEnt.samples, "/Sp290.csv", "
>> outputdirectory=", MaxEnt.out, "  maximumiterations=100 redoifexists autorun
>> nowarnings notooltips", sep=""))
>> # NOT RUN
>> system(command=paste("java -mx4000m -jar maxent.jar environmentallayers=",
>> MaxEnt.layers, " samplesfile=", MaxEnt.samples, "/Sp290.csv", "
>> outputdirectory=", MaxEnt.out, " 'applythresholdrule=equal training
>> sensitivity and specificity' maximumiterations=100 redoifexists autorun
>> nowarnings notooltips", sep=""))
>> list.files(MaxEnt.out)
>> #Import the predictions on the presence points
>> Pred.Maxent.Sp290.sample <- read.csv(paste(getwd(),
>> "/MaxEnt.res/Sp290_samplePredictions.csv", sep=""))
>> #Plot the results from Maxent
>> library(BIOMOD)
>> par(mfrow=c(1,2))
>> level.plot(Sp.Env$Sp290, XY=Sp.Env[,2:3], color.gradient = "grey",
>> cex=0.3,show.scale=F, title="Original data")
>> level.plot(Pred.Maxent.Sp290.sample$Logistic.prediction,
>> XY=Pred.Maxent.Sp290.sample[,1:2], color.gradient = "grey",
>> cex=0.3,show.scale=F, title="Maxent")
>> points(Sp.Env[Sp.Env$Sp290<1,2], Sp.Env[Sp.Env$Sp290<1,3], cex=0.3,
>> pch=2,col="grey93")
>> 
>> You see, this is not that difficult. I just need to see how to fully
>> implement this in the existing structure of BIOMOD, the most difficult part
>> obviously.
>> Cheers,
>> Wilfried
>> 
>> 
>> 
>> 
>> Thanks for the attention,
>> 
>> Best regards,
>> 
>> Charles
>> 
>> --
>> Um axé! :)
>> 
>> --
>> Charles Novaes de Santana
>> PhD student - Global Change
>> Laboratorio Internacional de Cambio Global
>> Department of Global Change Research
>> Instituto Mediterráneo de Estudios Avanzados(CSIC/UIB)
>> Calle Miquel Marques 21, 07006
>> Esporles - Islas Baleares - España
>> _______________________________________________
>> 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
>> Personal website: http://www.will.chez-alice.fr
>> Team website: http://www-leca.ujf-grenoble.fr/equipes/emabio.htm
>> 
>> FP6 European MACIS project: http://www.macis-project.net
>> FP6 European EcoChange project: http://www.ecochange-project.eu
>> 
>> 
>> 
>> 
>> 
>> 
> 
> 
> 
> -- 
> Um axé! :)
> 
> --
> Charles Novaes de Santana
> PhD student - Global Change
> Laboratorio Internacional de Cambio Global
> Department of Global Change Research
> Instituto Mediterráneo de Estudios Avanzados(CSIC/UIB)
> Calle Miquel Marques 21, 07006
> Esporles - Islas Baleares - España
> _______________________________________________
> 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
Personal website: http://www.will.chez-alice.fr
Team website: http://www-leca.ujf-grenoble.fr/equipes/emabio.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/20110426/1965605a/attachment-0001.htm>