Function to calculate and format posterior prediction distributions from INLA using random error trick

One way to calculate the PI for an INLA model with the appropriate error is to fit a random error term to each observation. This takes the output of that to estimate PIs.

Usage

get_inla_pi(inla_marginal, quantile_levels, id_df)

Arguments

inla_marginal

output of marginals.fitted.values from INLA model for one observation

quantile_levels

quantiles that you want to estimate interval at

id_df

a one row dataframe containing the id variables and true_value for that observation