brms_inla_power_sequential: Sequential Bayesian Assurance Simulation Engine (Modern,...

In powerbrmsINLA: Bayesian Power Analysis Using 'brms' and 'INLA'

Sequential Bayesian Assurance Simulation Engine (Modern, Multi-Effect Ready)

Description

Simulates assurance sequentially in batches, stopping early per cell based on Wilson confidence intervals.

Usage

brms_inla_power_sequential(
  formula,
  family = gaussian(),
  family_control = NULL,
  Ntrials = NULL,
  E = NULL,
  scale = NULL,
  priors = NULL,
  data_generator = NULL,
  effect_name,
  effect_grid,
  sample_sizes,
  metric = c("direction", "threshold", "rope", "bf"),
  target = 0.8,
  prob_threshold = 0.95,
  effect_threshold = 0,
  rope_bounds = NULL,
  credible_level = 0.95,
  compute_bayes_factor = FALSE,
  error_sd = 1,
  group_sd = 0.5,
  obs_per_group = 10,
  predictor_means = NULL,
  predictor_sds = NULL,
  seed = 123,
  batch_size = 20,
  min_sims = 40,
  max_sims = 600,
  ci_conf = 0.95,
  margin = 0.02,
  inla_num_threads = NULL,
  family_args = list(),
  progress = TRUE
)

Arguments

formula	brms-style model formula.
family	GLM family (e.g., gaussian(), binomial()).
family_control	Optional list for INLA's control.family.
Ntrials	Optional vector of binomial trial counts (for binomial families).
E	Optional vector of exposures (for Poisson families).
scale	Optional numeric vector for scale parameter in INLA.
priors	brms prior specification object. Supported priors are translated to INLA controls where possible and audited in settings$prior_translation.
data_generator	Optional function(n, effect) to simulate data.
effect_name	Character vector of fixed effects to assess.
effect_grid	Data frame or vector of effect values.
sample_sizes	Vector of sample sizes.
metric	Character; one of "direction", "threshold", "rope", or "bf" for Bayesian decision metric.
target	Target conditional power for the stopping rule (0-1).
prob_threshold	Posterior probability threshold for decision metrics.
effect_threshold	Effect-size threshold.
rope_bounds	Numeric length-2 vector defining ROPE.
credible_level	Credible interval level for Bayesian inference.
compute_bayes_factor	Logical; TRUE if metric is "bf".
error_sd	Residual standard deviation.
group_sd	Standard deviation of random effects.
obs_per_group	Number of observations per group.
predictor_means	Optional named list of predictor means.
predictor_sds	Optional named list of predictor standard deviations.
seed	Random seed.
batch_size	Number of simulations per sequential look.
min_sims	Minimum simulations before early stopping.
max_sims	Maximum simulations per cell.
ci_conf	Confidence level for Wilson confidence intervals.
margin	Margin around target for early stopping decision.
inla_num_threads	Character string specifying INLA threading (e.g., "4:1"). If NULL (default), automatically detects optimal setting based on CPU cores.
family_args	List of family-specific args passed to data generator.
progress	Logical; if TRUE, show progress messages.

Details

Sequential Bayesian Assurance Simulation Engine (Modern, Multi-Effect Ready)

Simulates assurance sequentially in batches, stopping early per cell based on Wilson confidence intervals.

Value

A list of class "brms_inla_power" with a per-cell summary (including the conditional_power column, the Monte Carlo estimate at each fixed effect value) and simulation settings. Note this is conditional power at each grid value, not unconditional assurance; pass the result to compute_assurance() for the latter.

Examples

## Not run: 
# Sequential design with automatic threading
results <- brms_inla_power_sequential(
  formula = outcome ~ treatment,
  effect_name = "treatment",
  effect_grid = c(0.2, 0.5, 0.8),
  sample_sizes = c(50, 100, 200),
  metric = "direction",
  target = 0.80
)
print(results$summary)

## End(Not run)

powerbrmsINLA documentation built on July 2, 2026, 5:07 p.m.