4.2 Inla()

Keys argument:

  • formula

  • data

  • family : type of likelihhod distributon we want to use as model

(inla.models()$likelihood)$poisson
## $doc
## [1] "The Poisson likelihood"
## 
## $hyper
## list()
## 
## $survival
## [1] FALSE
## 
## $discrete
## [1] TRUE
## 
## $link
## [1] "default"   "log"       "logoffset" "quantile"  "test1"     "special1" 
## [7] "special2" 
## 
## $pdf
## [1] "poisson"
#inla.doc("poisson") # replace poisson with other family

  • control.compute : take a list with “key = value” of other variable that should be computed. Example with deviance information criterion (DIC)

  • control.predictor : take a list that specify predictor variable (link) but also if marginal densities of LP should be computed