Skip to contents

Returns the prior that rstanarm placed on each parameter of a rstanarm::stan_glm() fit, after autoscaling. For rstanarm::stan_glmer() fits, the population-level parameters are returned; the group-level covariance prior is described by rstanarm::prior_summary(fit)$prior_covariance. rstanarm's rstanarm::prior_summary() prints the same information; this function returns it as a data frame.

Usage

prior_table(fit)

Arguments

fit

A stanreg object from rstanarm::stan_glm(), rstanarm::stan_glmer() or fit_model_with_prior(), fitted without QR = TRUE.

Value

A data frame with one row per parameter and columns variable, role ("intercept", "slope" or "aux"), dist, location, scale, df, prior_sd (the prior standard deviation, Inf when it does not exist) and prior (a readable description).

Details

The intercept row describes the prior on the intercept at the predictor means, which is the parameter rstanarm places its prior on. Auxiliary parameters (e.g. sigma) have half-distributions on the positive real line, except for the exponential.

Examples

# \donttest{
data(kidiq, package = "rstanarm")
fit <- rstanarm::stan_glm(kid_score ~ mom_iq, data = kidiq,
                          chains = 2, iter = 1000, refresh = 0)
prior_table(fit)
#>      variable      role        dist location     scale df  prior_sd
#> 1 (Intercept) intercept      normal 86.79724 51.026721 NA 51.026721
#> 2      mom_iq     slope      normal  0.00000  3.401781 NA  3.401781
#> 3       sigma       aux exponential       NA 20.410688 NA 20.410688
#>                       prior
#> 1          normal(86.8, 51)
#> 2            normal(0, 3.4)
#> 3 exponential(rate = 0.049)
# }