Issues with Precision Parameters Using INLA for Bivariate Spatial Poisson Regression with Shared Component

[mahsa nadifar]

  Hi INLA support team,

I am conducting a simulation study using a bivariate spatial Poisson regression model with a shared component. I have fitted the model using both INLA and OpenBUGS, but I am encountering significant differences in the precision parameters between the two methods. The precision parameters estimated by INLA are quite different from the actual values.

Here is my code for reference.

I was wondering if you could provide insights on what might be causing this difference.

I appreciate your help!

Best regards,

Mahsa,  

[INLA help]

I think you forgot to add E1 and E2 in the inla call 

—

Haavard Rue

Helpdesk

he...@r-inla.org

--
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/5f1fc13b-600c-4a9e-b4a6-051af96cc860n%40googlegroups.com.

[mahsa nadifar]

Hi,  

Thank you for your response. I have added E1 and E2 to the model, but the results are still different!

# offset

E = c(E1, E2)


formula <- Y ~ -1 + I1 + I2 +
  f(IDs1, model = "bym", graph = W, scale.model=TRUE, constr=TRUE) +
  f(IDs2, copy = "IDs1", fixed = FALSE,
hyper = list(beta = list(prior = 'normal', param = c(0, 10))),
range=c(0, Inf)) +
  f(IDs1.2, model = "bym", graph = W, scale.model=TRUE, constr=TRUE) +
  f(IDs2.2, model = "bym", graph = W, scale.model=TRUE, constr=TRUE)



res1 <- inla(formula = formula, data = D, family = rep("poisson", K), E=E,
control.compute=list(dic=TRUE, waic=TRUE, cpo=TRUE))

[INLA help]

And the priors are the same ?

—

Haavard Rue

Helpdesk

he...@r-inla.org

--

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/babd5197-3b18-4448-95d9-eb805d93d155n%40googlegroups.com.

[mahsa nadifar]

Yes, as a default of precision prior in INLA, I considered gamma(1,  5e-4)  for precision parameters.

[Helpdesk (Haavard Rue)]

I take this thread off the list as its getting to spesific

> https://groups.google.com/d/msgid/r-inla-discussion-group/7464d258-b96a-4182-b927-8080532ea910n%40googlegroups.com
> .

--
Håvard Rue
he...@r-inla.org