Is currently not in a reliable state useful for testing. Needs more thought.
Usage
test_rng_asym(
rng_fun,
metric_mu,
n_samples = c(10, 10000),
mu_list,
aux_list,
aux2_list = NA,
mu_link = identity,
allowed_failures = 0.05
)Arguments
- rng_fun
RNG function under test
- metric_mu
Metric to be used on RNG data (usually mean or median)
- n_samples
Default=c(10, 10000), sample sizes for rng test, len >= 2. Gets sorted in routine.
- 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 for second parameter (if applicable).
- mu_link
Default=identity, optional link-function argument, for example
- allowed_failures
Default=0.05 marks, that 5 percent of all test cases are allowed to fail for the whole test to still succeed.
Examples
eps <- 0.001
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_asym(
rng_fun = rbetaprime,
metric_mu = mean,
mu_list = mu_list,
aux_list = phis,
)
print(result)
#> <expectation_success/expectation/condition>
#> As expected