Species distribution modeling in INLA

[gachie thomas]

Hello

I am building a habitat suitability model (species distribution model) for a particular mosquito species. 

Below is a brief description of my data.

1. data on the presence of the mosquito species in a particular location is collected from different sources.
   
2. pseudo absence points are also generated.
   

using known climatic conditions known to influence the habitat of the species, i want to build a presence absence (0/1) suitability model in INLA.

I have been able to do that using different machine learning algorithms, wondered if its possible to build such kind of a model in INLA. If its possible, how would i specify my model.

Thank you.

Thomas.

[Marie-Christine Rufener]

Hi Thomas,

You can do SDMs perfectly well through INLA (I have done it already several times.)
As far as I know, regarding specifically Bayesian geostatistical models, the first published paper applying this kind of idea through INLA was those of Muñoz et al. (see paper Paper 1). After that, it has been increasingly used in fisheries science. Mrs. Maria G. Pennino has been applying it constantly (see papers on SMEG), and me since a while too. For further methodological/theoretical details about its application in SDMs, you may refer to Paper 1, supplementary material of Paper 2, and/or Paper 3.

I have included a detailed script in my master thesis about how to perform SDMs in INLA (unfortunately it is in Portuguese, but in case you want, I may translate it in english). 
Also, feel free to contact Mrs. Pennino (or me) for further questions, or of course, the own r-inla group! 

Best wishes,

Marie

[Marie-Christine Rufener]

Ps.: I forgot to mention, but in all cases mentioned before, INLA was exclusively used to perform geostatistical models - not machine learning algorithms! In this case, unfortunately I have no idea how one could adapt machine learning algorithms to INLA.

On Thursday, April 13, 2017 at 5:31:18 PM UTC+2, gachie thomas wrote:

[Bob O'Hara]

I'm developing a package to fit SDMs to different types of data,
including presence-only. Presence-only data is probably best modelled
as a point process ( e.g. see
http://onlinelibrary.wiley.com/doi/10.1111/2041-210X.12352/abstract),
which INLA handles really well: pseudo-absences are really just points
being used in a numeric integration, and the INLA crowd showed how to
do this efficiently. The SPDE tutorial
(http://www.r-inla.org/examples/tutorials) is a good starting point.

For the different machine learning models, as long a they can be
written as linear models, they can be fitted with INLA. There is now a
MaxNet package in R which can probably be used to get the MaxEnt
bases, if you want to poke around a bit.

Hm, I'm not sure if this is a huge help - email me if you want to know
more about my package. It's not quite ready to release into the wild,
but is close.

Bob

> --
> You received this message because you are subscribed to the Google Groups
> "R-inla discussion group" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to r-inla-discussion...@googlegroups.com.
> To post to this group, send email to
> r-inla-disc...@googlegroups.com.
> Visit this group at https://groups.google.com/group/r-inla-discussion-group.
> For more options, visit https://groups.google.com/d/optout.



--
Bob O'Hara

Biodiversity and Climate Research Centre
Senckenberganlage 25
D-60325 Frankfurt am Main,
Germany

Tel: +49 69 798 40226
Mobile: +49 1515 888 5440
WWW: http://www.bik-f.de/root/index.php?page_id=219
Blog: http://occamstypewriter.org/boboh/
Journal of Negative Results - EEB: www.jnr-eeb.org

[INLA help]

On Thu, 2017-04-13 at 19:52 +0200, Bob O'Hara wrote:
>
> For the different machine learning models, as long a they can be
> written as linear models, they can be fitted with INLA. There is now
> a
> MaxNet package in R which can probably be used to get the MaxEnt
> bases, if you want to poke around a bit.
>

for a such comparison/discussion, see

http://biorxiv.org/content/early/2017/02/06/105742



H

--
Håvard Rue
he...@r-inla.org