You can find a ppt on the ENFA/Biomapper website, explaining how to choose thresholds using the Boyce-index. The presentation is "Evaluation of presence-only models with the continuous Boyce index and cross-validation. Presented at a workshop at the Castle of Rauischholzhausen (Germany) in 2006. (30 minutes)", and can be found here: http://www2.unil.ch/biomapper/Presentations.html

An relevant paper is: Hirzel, A., Le Lay, G., Helfer, V., Randin, C. & Guisan, A. (2006) Evaluating the ability of habitat suitability models to predict species presences. Ecol Model, 199, 142–152.

Cheers,

Diederk

Thank you Diederk

i'll give a look!!

Ciao Lorenzo. Ho girato la tua domanda a Mirko di Febbraro, il modellista che mi aiuta. Ecco la sua risposta: 

Ciao Laura,

ho letto la domanda e mi pare interessante. Ora come ora non ho idea di come poter fare. Ci penso e cerco un po' su internet. Se trovo qualcosa te lo faccio sapere.

Ti chiederei anzi il favore in direzione opposta: se il tuo collega trova materiale interessante, sarei contento di poterlo avere.

Grazie.

A presto

M

ciao Laura,

grazie per l'interessamento e per avermi girato la risposta.

Se ho novità mi faccio vivo...

ciao 

Lorenzo

Hi Lorenzo Lazzaro ,

Did you find the answer to your question, if so then could you please kindly share how to do it; step-wise or if you have R code. It would be of great help because I am also stuck at the same question because i used Maxent GUI with cross-validation so basically I dont have evaluation dataset set aside as it varies for each replicates (10 in my case). So, I am thinking how do I do it?

Hoping to hear from you.

Regards

Sujata