Skip to contents

Convergence diagnostics of a fitted model

Source: R/report_methods.R
report_convergence.Rd

Returns the convergence diagnostics of a model fitted with bmm(), per parameter or as the sentence that report_methods() uses.

Usage

report_convergence(fit, format = c("table", "text"))

Arguments

fit

A bmmfit object returned by bmm(). It must contain posterior draws, so mock fits and fits made with chains = 0 or empty = TRUE are refused.

format

"table" for one row per parameter, "text" for the sentence.

Value

An object of class bmm_report_convergence: a data.frame with columns parameter, rhat, ess_bulk, ess_tail and in_summary for format = "table", a character string for format = "text". Subsetting the table with [ returns a plain data.frame.

Details

The table holds the rank-normalized R-hat and the bulk and tail effective sample sizes (ESS) of every estimated parameter, computed with posterior::summarise_draws(). Constants, lp__, lprior and the draws from the prior (prior_*) are left out. The column in_summary marks the parameters that summary() shows; brms warns when one of them has an R-hat above 1.05. The algorithm, the number of divergent transitions after warmup and the number of transitions that reached the maximum tree depth are attributes of the table, printed below it.

The sentence gives the largest R-hat, the smallest bulk and tail ESS and the number of divergent transitions, without verdicts. For algorithms other than "sampling" the table is empty and the text says that there are no diagnostics.

See also

report_methods()

Examples

if (FALSE) { # identical(Sys.getenv("BMM_EXAMPLES"), "true")
fit <- bmm(
  bmf(c ~ 0 + set_size, kappa ~ 1),
  data = oberauer_lin_2017,
  model = sdm(resp_error = "dev_rad")
)
report_convergence(fit)
report_convergence(fit, format = "text")
}