Log Gaussian Cox Process

[houessou madjid]

Hi all,

I am very new to INLA, and I am looking at some examples of using INLA for fitting spatial-temporal Log Gaussian Cox Process model.

Concretely, I have a bird citizen science dataset and some spatiotemporal environmental variable from daymet. 

I would like to set up a spatiotemporal Log Gaussian Cox Process which account for spatio-temporal dependencies.

My model is defined  like this : 

\ln(\lambda(s_i,t)) &= X(s_i,t)\beta+\{Z(s) \otimes Z(t)\}^T u 

with with $u \sim \mathcal{N}(0,\sigma^{2}\mathbf{I})$. Z(s) and Z(t) are spatial and temporal basis functions.

 Is it possible to implement it in INLA? If yes, any example will be very appreciate.

Thanks,

[Finn Lindgren]

As a starting point, take a look at the inlabru examples, in particular the space-time example: (please don’t try to set this up in plain INLA; inlabru exists to handle these types of model and make it easier to focus on the model rather than low-level model construction details)

Finn

On 15 Jul 2025, at 21:51, houessou madjid <hmad...@gmail.com> wrote:

Hi all,

--
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 view this discussion, visit https://groups.google.com/d/msgid/r-inla-discussion-group/b5b14b0a-de8f-4a7d-bc9f-a92aadb72cf9n%40googlegroups.com.

[Finn Lindgren]

Note, for space-time covariates stored as a terra raster with each time point as a layer, there is a bug in the terra package that can cause the wrong values to be extracted when using numerical layer indices, see

As a workaround, convert the layer indices to layer names when doing the indexing.

Something like this:

names(mycovar) <- as.character(seq_len(nlyr(mycovar)))

In the component definitions,

+ thecovar(eval_spatial(mycovar, .data., layer =as.character( time)), model=“linear”)

Finn

On 15 Jul 2025, at 23:39, Finn Lindgren <finn.l...@gmail.com> wrote:

As a starting point, take a look at the inlabru examples, in particular the space-time example: (please don’t try to set this up in plain INLA; inlabru exists to handle these types of model and make it easier to focus on the model rather than low-level model construction details)