Model replication

[alaaeldeen80]

Dear List,

I'm using maxent to model my species with few presence points 9 without data splitting and with 100 replications but I found that maxent ran only 9 replications.
I did try to change my presence points for example i used 5 , 6, 7 ,8 presence points and in each time it ran the model with number of replications equal to the number of presence points and not with 100 replication as I adjust the maxent setting.

Any thought will be appreciated

Alaa

[Husam El Alqamy]

Dear Alaa

Replication is subsetting your full dataset at random each time. Since you have only ten points only few replications are possible which means that maxent was randomizing within 10% of your sample size every time until it ran out of possible combinations.

Regards

 

Husam El Alqamy, M.Phil.

Sr. Biodiversity GIS Analyst

Environment Agency – Abu Dhabi

Antelope Specialist Group, IUCN

Research Profile

--
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 http://groups.google.com/group/maxent.
For more options, visit https://groups.google.com/d/optout.

[Alaa Eldeen]

Thanks Husam for your reply.

i just want to know if I would like to run my model with many replications and in each replication I will use different background points, Up to my understanding, I will get different result so I will use the mean to get the mean of all replications.
Am I right?

Cheers

--

Alaa Eldin Soultan,  B.Sc, MSc.
Wildlife Conservationist,
St Katherine Protectorate,
National parks of Egypt,
Nature Conservation Sector,
Egyptian Environmental Affairs Agency, EEAA
Tel-Fax: + 2069- 3470033
Mobil:+201225773922



Ps:  It takes 24 trees to produce 1 ton of office paper! Think… is it really necessary to print this email?

[Husam El Alqamy]

Dear Alaa

How different are the models depends on the distribution of your sample points and how variable are the environmental data across the study region. If there are big difference among yo9ur replica models it would be hard to average them in a one representative model.

[Ahmed El-Gabbas]

Dear Alaa,

The Maxent's default way of dealing with replications is via cross-validation 'CV'. As you have only 9 points, there is no more than 9 cross-validation runs possible (using the leave-one-out approach: 'jacknifing', the most extreme case of CV), whatever how many replicates you explicitly specified in 'Replicates' option : 100 in your case.

In your case, and using other default options, you will have just 9 different models, each calibrated with 8 different combinations of points and leave one point aside for testing. This will be done iteratively, and in each model a different one point is used for testing. I suspect that maxent uses the same random 10,000 background points at each of the sub-models (unless you check the 'random seed' option, see below), which is not what you need.

I think you need to check 'random seed' option to allow different random background points to be sampled from the landscape at each of the 100 runs AND not to choose cross-validation option as "replicated run type". Instead, I suggest you to use the "Subsample" option along with keeping the 'random test percentage' option = 0. I think this way, the number of models to be run will not be constrained by maxent and you should have 100 different models, each uses the same 9 points for calibration and different (if your landscape allows) 10,000 background points.
You can also consider using the "Bootstrap" method: Sampling with replacement, or changing the 'random test percentage' option depending on the research question you have.

--------
Regards,
Ahmed ElGabbas

[Alaa Eldeen]

Dear Ahmed,

Thanks for clarification, yes this was my original thought about using CV or sub-sampling.

Ramdan Kareem

--

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 http://groups.google.com/group/maxent.
For more options, visit https://groups.google.com/d/optout.