Apply SpANOVA modelization using a wrapper function in INLA

This function is a wrapper for multidimensional spatial factor models in INLA, using a sequential shared spatial effects with nested structure as discussed in AÑADIR REFERENCIA.

Usage

inla.SpANOVA.2x2(
  data,
  gr,
  fac.names = NULL,
  lev.fac1 = NULL,
  lev.fac2 = NULL,
  scale.mod = TRUE,
  sp.prior = "sdunif",
  pc.prec.val = c(1, 0.01),
  sp.copy.fixed = TRUE,
  save.res = FALSE,
  save.random = TRUE,
  save.hyper = TRUE,
  save.fixed = TRUE,
  save.mod.data = FALSE,
  verbose.INLA = FALSE
)

Arguments

data

Dataframe containing the number of events observed on column obs, the expected values on column exp, the level for factor 1 on column lev.fac1 and the level for factor 1 on column lev.fac2

gr

Graph for the underlying spatial structure

fac.names

Names of the factors included

lev.fac1

Levels of the first factor included

lev.fac2

Levels of the second factor included

scale.mod

Scale copies of random spatial effects or not, default is TRUE

sp.prior

Select prior for the random spatial effect, options are sdunif and pc.prec

pc.prec.val

Define values por the pc prior in case it was chosen. Default values are c(1, 0.01)

sp.copy.fixed

Fix copied values for the random spatial effects, default is TRUE

save.res

Save fitted values or not from the different models, default is TRUE

save.random

Save values adjusted from the random spatial effects or not from the different models, default is TRUE

save.hyper

Save hyperparameter values from each individual model, default is TRUE

save.fixed

Save values adjusted from the fixed effects or not from the different models, default is TRUE

save.mod.data

Save modelling data to run the model afterwards

verbose.INLA

Verbose option for INLA, default is FALSE

Value

List with all the models analyzed and a summary table with the most common performance metrics.