STb_compare_lfo() — STb_compare

Compare two or more objects returned by STb_lfo() using leave-future-out expected log predictive density. The model with the highest total elpd_lfo is treated as the reference model. Differences are reported relative to that model.

Usage

STb_compare_lfo(..., model_names = NULL, n_obs = NULL, digits = 1)

Arguments

...: Two or more LFO objects returned by STb_lfo().
model_names: Optional character vector of model names. If NULL, names are taken from ... when available, otherwise generic names are used.
n_obs: Optional integer giving the number of pointwise observations to use when computing standard errors. If NULL, the number of complete pointwise observations across models is used.
digits: Integer. Number of digits used for rounding the returned table. Defaults to 1.

Value

A data frame with columns:

elpd_diff: Difference in total ELPD from the best model.
se_diff: Standard error of the ELPD difference.
elpd_lfo: Total leave-future-out ELPD.
se_elpd_lfo: Standard error of total leave-future-out ELPD.
lfoic: LFO information criterion, computed as -2 * elpd_lfo.
se_lfoic: Standard error of lfoic.