Case: housing prices in Boston

Fitting the INLA model:

res <- inla(formula, family = "gaussian", data = map,
            control.predictor = list(compute = TRUE),
            control.compute = list(return.marginals.predictor = TRUE))

Control computation of fitted values (at the scale of the linear predictor):

• control.predictor = list(compute = TRUE): fitted values and their posterior summary: res$summary.fitted.values
• control.compute = list(return.marginals.predictor = TRUE): compute posterior marginal distribution of the fitted values: res$marginals.fitted.values[[1]]

Model object evaluation: see book.