Creates a formal S3 object containing posterior samples and metadata from Bayesian pooling model fitting (EXNEX, complete pooling, or no pooling).
Usage
new_pooling_surv(
draws,
data,
pooling,
priors,
resolved_priors = priors,
iter,
warmup,
thin = 1L,
chains,
blueprint
)Arguments
- draws
A data frame containing posterior samples. Columns are: theta_1, ..., theta_K, beta_1, ..., beta_P (if P > 0), sigma2.
- data
A list containing: - time: Observed follow-up times - event: Event indicators (0/1) - group: Group assignments - X: Covariate matrix (can be empty) - n: Total number of observations - n_groups: Number of groups - n_covariates: Number of covariates - cov_names: Covariate column names
- pooling
The fitted pooling variant: one of
"exnex","complete", or"none".- priors
A list of prior specifications used for fitting.
- resolved_priors
A named list of the prior hyperparameters actually used (defaults merged with any overrides supplied in
priors).- iter
Total MCMC iterations performed.
- warmup
Number of warmup iterations discarded.
- thin
Thinning interval applied to the post-warmup draws.
- chains
Number of chains run.
- blueprint
The hardhat blueprint for the original formula/data structure.