Guidance on kappa and TSS values for Maxent output..

[Asad Ali]

Hi all,

I have been using maxent jar software for generating background points for my true presences. I actually am a geostatistician and wanted to apply kriging prediction method to generate prediction maps for certain species. However, I realized that with maxent, the best number of background points is 10,000 as argued by many authors. Now 10000 backgound points plus actual presences make it a huge number, not feasible for distance based methods such as kriging. It's hard to get a sense of true absence probability due to multiple reasons.  So I started experimenting on the number of background points. I generated multiple datasets starting from a small number of background points taking it to 10,000 gradually. For example for a data of 386 presences I used 15, 30, 50, 100, 200, 300, 386(equal), 400, 500, ....., 2000, ...., 3000, ....5000,..., 10000. I had to look at the geostatistical properties of the spatial structure (variogram) and maps of each dataset.  I realized that only a moderate number  of background points (equivalent to number of presences) gives the best spatial structure modeling. My colleague then helped me prepare a table of AUCs, Kappa and TSS, for all datasets, in which dataset with equal number of presences and background points gave the highest kappa and TSS values, supporting my geostatistical results of stronger spatial structure.

What I am asking here is "am I doing it correctly"? Are my findings make sense?

Also, is there any R code which can compute both Kappa and TSS from Maxent output?

Cheers,

Asad   

PS: I have been fulfilling all the requirements for Maxent, such as bias file, environmental predictors etc. 

[Alaaeldin Soultan]

Hi Asad

I would not recommending using Kappa since you are using background not the true absence data.

Have a look to this paper it might help you

https://besjournals.onlinelibrary.wiley.com/doi/pdf/10.1111/j.2041-210X.2011.00172.x

there are a  lot of r packages(e.g. SDMTools) that allow for calculating TSS and Kappa

I hope it helps

--
You received this message because you are subscribed to the Google Groups "Maxent" group.
To unsubscribe from this group and stop receiving emails from it, 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.

--

Alaaeldin Soultan
PhD candidate
MaxPlanck Institute for Ornithology
Am Obstberg 1
78315 Radolfzell, Germany
Tel: +49 (0)7732 1501 35
Mob.+4915757474779