Check LOO Diagnostics for brma Objects — check

Check Pareto k diagnostics for a brma model object and warn if any values are high.

Usage

# S3 method for class 'brma'
check_loo(object, unit = "estimate", ...)

Arguments

object: a brma model object.
unit: output/deletion unit. See add_loo.
...: currently unused.

Value

NULL (throws warning if diagnostics are unreliable).

Details

LOO must first be computed with object <- add_loo(object, unit = unit). With \(S\) posterior draws in the stored LOO result, the method warns when any Pareto \(k\) diagnostic is greater than \(\min(1 - 1 / \log_{10}(S), 0.7)\). The warning gives that cutoff and the affected target indices.