Tests if an RNG can recover the true quantiles within a margin of error

Usage

test_rng_quantiles(
  rng_fun,
  quantile_fun,
  n,
  mu_list,
  aux_list = NA,
  aux2_list = NA,
  eps,
  quantiles,
  p_acceptable_failures,
  mu_link = identity,
  relative = FALSE
)

Arguments

rng_fun: RNG function under test
quantile_fun: Quantile function related to the rng under test.
n: Sample size for the rng test.
mu_list: Metric data used as RNG argument and to be compared to (usually mean or median)
aux_list: Auxiliary parameter value list.
aux2_list: Auxiliary parameter value list for applicable distributions.
eps: Acceptable difference of |mu - metric_mu(rng_fun)
quantiles: Quantiles to test for recovery.
p_acceptable_failures: Acceptable rate of failure, relative value of difference bigger mu_eps
mu_link: Default=identity, optional link-function argument, for example useful in link-normal-distributions
relative: True if the error should be relative to the mu_list

Value

Nothing actually, just wraps the test

Examples

eps <- 0.001
mu_list <- seq(from = 1 + eps, to = 20, length.out = 10)
phi_list <- seq(from = 2 + eps, to = 20, length.out = 10)
# if working as expected, this test should not print any errors
bayesfam:::test_rng_quantiles(
  rng_fun = rbetaprime,
  quantile_fun = qbetaprime,
  n = 10000,
  mu_list = mu_list,
  aux_list = phi_list,
  eps = 0.1,
  quantiles = c(0.01, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 0.95, 0.99),
  p_acceptable_failures = 0.1,
  relative = TRUE
)