About plotting exceedance probability from a spatio-temporal model

[GLORIA NKEM IBEH]

I fitted spatio-temporal model and wanted to plot the exceedance probability but I found out that the list of  the marginal posterior distribution of fitted values   "marginals.fitted.values" is returning NULL value, how can I resolve this?  I have also tried to work with the  marginals of the linear predictor​, transform them to the fitted value yet it did not work. Can anyone help me out on this

 here is my model:                                                                                                                               #Type 1(interaction between two unstructured effect)
formula.hosp.rw2.1A <- Incidence ~ 1 + CUMUL_vacc_rate*stringency_index+ logCASES +
f(provID, model="bym2",graph=grp,scale.model=TRUE) +
f(week_ID,model="rw2") +
f(week_id.1,model="iid") +
f(ID.prov.week,model="iid")  
covid.hosp.rw2.1A <- inla(formula.hosp.rw2.1A,family="poisson",
        data=data.set,E= expected_hosp,
        control.predictor=list(compute=TRUE, link =1), #enable prediction
        control.compute=list(dic=TRUE, cpo=TRUE),
        lincomb=lcs)
exc<-sapply(covid.hosp.rw2.1A$marginals.fitted.values,
 FUN =function(marg){1-inla.pmarginal(q =1,marginal =marg)})           

[Johnson, Daniel P]

control.compute = list(config = TRUE)

Daniel P. Johnson, Ph.D.

Associate Professor

Indiana University at Indianapolis
Department of Geography

dpjo...@iu.edu

317-278-5536

---

From: r-inla-disc...@googlegroups.com <r-inla-disc...@googlegroups.com> on behalf of GLORIA NKEM IBEH <ibehg...@gmail.com>
Sent: Wednesday, August 13, 2025 12:28 PM
To: R-inla discussion group <r-inla-disc...@googlegroups.com>
Subject: [External] [r-inla] About plotting exceedance probability from a spatio-temporal model

 

You don't often get email from ibehg...@gmail.com. Learn why this is important

This message was sent from a non-IU address. Please exercise caution when clicking links or opening attachments from external sources.

--
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/436f7946-769f-4d8c-9146-625f7ac05823n%40googlegroups.com.

[GLORIA NKEM IBEH]

Thank you  Johnson, Daniel P for your insight however, I tried it yet it did not work.

[Helpdesk (Haavard Rue)]

Note that the value of 'E' is not part of the fitted.values (its a part of the
likelihood...), but you can move 'E' into the linear predictor using argument

offset = log(expected_hosp)

instead of

E = expected_hosp

or add offset(log(expected_hosp)) to the formula

> y=10
> r=inla(y~1,data=data.frame(y), family="poisson")
> r$summary.fitted.values$mean
[1] 9.999991924
> r=inla(y~1,E=10, data=data.frame(y), family="poisson")
> r$summary.fitted.values$mean
[1] 0.9999991925
> r=inla(y~1,offset=log(10), data=data.frame(y), family="poisson")
> r$summary.fitted.values$mean
[1] 9.99998304
> r=inla(y~1 + offset(log(10)), data=data.frame(y), family="poisson")
> r$summary.fitted.values$mean
[1] 9.99998304

On Wed, 2025-08-13 at 11:48 -0700, GLORIA NKEM IBEH wrote:
> Thank you Johnson, Daniel P for your insight however, I tried it yet it did
> not work.
>
> On Wednesday, August 13, 2025 at 6:04:40 PM UTC+1 Johnson, Daniel P wrote:
> >
> >
> > control.compute = list(config = TRUE)
> >
> >
> >
> >
> > Daniel P. Johnson, Ph.D.
> > Associate Professor
> > Indiana University at Indianapolis
> > Department of Geography
> >
> > dpjo...@iu.edu
> >
> > 317-278-5536
> >
> >
> >
> >
> >

Håvard Rue
he...@r-inla.org