Interpretation of BLUPs

[Alexander Jack]

Hi all,

I have a model with several ar(1) processes. I am having trouble interpreting the output of model$summary.random$mean. I would like to plot these BLUPs against time and interpret them. The formula in question is:

f(as.factor(idtime3), scale(temp), model="ar1", constr=TRUE)

idtime3 is the year (1-15 in the output) and temp is a climate covariate I'm interested in. I would like help understanding the mean and the credible intervals around it from the output (attached). Some books have pointed me toward these as "smoothing" parameters, but ultimately what I want to know is some way to discuss how this variable changes with time...

[Helpdesk (Haavard Rue)]

you should not use 'as.factor(idtime3))' as argument, can you use just 

f(idtime3, scale(temp), model="ar1") 

instead ???

let me know if that change anything.

otherwise,

model$summary.random$idtime3$mean

will give the pointwise mean and sd, for the AR1 process at time 1, 2, ..., 15.
(if you have years missing, you need to add the argument f(..., values=1:15)



On Fri, 2024-07-19 at 11:51 -0700, 'Alexander Jack' via R-inla discussion group
wrote:

> --
> 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 on the web, visit
> https://groups.google.com/d/msgid/r-inla-discussion-group/787c5e68-e5e0-4fcd-9a38-709c67773459n%40googlegroups.com
> .

--
Håvard Rue
he...@r-inla.org

[Alexander Jack]

I guess the reason why I wanted to use it as a factor was because I didn't want to assume it was monotonically increasing. So if each y is the Delta in the regression coefficient what would the fixed effect for scale(temp) mean?

[Alexander Jack]

In the same regression as this random ar process