Skip to contents

Sample from prior predictive distribution

Source: R/prior_predictive.R
sample_prior_predictive.Rd

Generates samples from the prior predictive distribution for the hierarchical shrinkage model. Useful for prior elicitation and sensitivity analysis.

The generative process is:

  1. Sample mu from p(mu)

  2. Sample tau from p(tau)

  3. Sample theta_i ~ N(mu, tau) for each group i

Usage

sample_prior_predictive(
  hierarchical_priors,
  n_groups,
  n_draws = 1000,
  group_names = NULL
)

Arguments

hierarchical_priors

Named list with mu and tau distributional objects from the distributional package.

n_groups

Integer; number of subgroups (G).

n_draws

Integer; number of prior predictive samples to draw. Default 1000.

group_names

Optional character vector of length n_groups to label groups.

Value

A list with class "shrinkr_prior_pred" containing:

mu

Vector of mu draws

tau

Vector of tau draws

theta

Matrix of theta draws (n_draws x n_groups)

implied_range

Vector of ranges (max - min) of theta across groups for each draw

implied_sd

Vector of standard deviations of theta across groups for each draw

group_names

Group labels

n_draws

Number of draws

n_groups

Number of groups

priors

The hierarchical_priors specification used

See also

shrink for fitting the hierarchical model, plot.shrinkr_prior_pred for visualizing prior predictive samples

Examples

priors <- list(
  mu = distributional::dist_normal(0, 5),
  tau = distributional::dist_truncated(distributional::dist_normal(0, 1), lower = 0)
)
prior_pred <- sample_prior_predictive(priors, n_groups = 3, n_draws = 50)
median(prior_pred$implied_range)
#> [1] 0.9419793
head(as.data.frame(prior_pred))
#> # A tibble: 6 × 5
#>   .draw group   theta     mu   tau
#>   <int> <chr>   <dbl>  <dbl> <dbl>
#> 1     1 group1  0.326  0.305 0.305
#> 2     1 group2 -0.159  0.305 0.305
#> 3     1 group3  0.568  0.305 0.305
#> 4     2 group1 -3.65  -4.12  0.930
#> 5     2 group2 -4.45  -4.12  0.930
#> 6     2 group3 -4.57  -4.12  0.930