Tests if an RNG can achieve a good enough location often enough

Tests if an RNG can achieve a good enough location often enough

Usage

test_rng(
  rng_fun,
  metric_mu,
  n,
  mu_list,
  aux_list = NA,
  aux2_list = NA,
  mu_eps,
  p_acceptable_failures,
  mu_link = identity,
  relative = FALSE,
  debug = TRUE
)

Arguments

rng_fun

RNG function under test

metric_mu

Metric to be used on RNG data (usually mean or median)

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

aux2_list

Auxiliary parameter for second parameter (if applicable).

mu_eps

Acceptable difference of |mu - metric_mu(rng_fun)

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, Default = FALSE

debug

bool argument, if set will printout the difference, in case of failure. Default = FALSE

Value

Nothing actually, just wraps the test

Examples

eps <- 1e-6
mu_list <- seq(from = 1 + eps, to = 20, length.out = 10)
phis <- seq(from = 2 + eps, to = 20, length.out = 10)
result <- bayesfam:::test_rng(
  rng_fun = rbetaprime, metric_mu = mean, n = 10000, mu_list = mu_list,
  aux_list = phis, mu_eps = 0.2, p_acceptable_failures = 0.05
)
print(result)
#> <expectation_success/expectation/condition>
#> As expected