Run posterior sampling for a prepared one-arm bundle

run_mcmc_bundle_manual() is the explicit runner for objects created by build_nimble_bundle. It compiles the stored NIMBLE code, executes MCMC, and returns a "mixgpd_fit" object.

Usage

run_mcmc_bundle_manual(
  bundle,
  show_progress = TRUE,
  quiet = FALSE,
  parallel_chains = FALSE,
  workers = NULL,
  timing = FALSE,
  z_update_every = NULL
)

Arguments

bundle

A causalmixgpd_bundle from build_nimble_bundle().

show_progress

Logical; if TRUE, print step messages and render progress where supported.

quiet

Logical; if TRUE, suppress console status messages. Set to FALSE to see progress messages during MCMC setup and execution.

parallel_chains

Logical; run chains concurrently when nchains > 1.

workers

Optional integer number of workers for parallel execution.

timing

Logical; if TRUE, include stage timings (build, compile, mcmc) in fit$timing.

z_update_every

Integer >= 1 controlling latent cluster-label update cadence.

Value

A fitted object of class "mixgpd_fit" containing posterior draws, model metadata, and cached objects used by downstream S3 methods.

Details

The resulting fit supports posterior summaries of the model parameters as well as posterior predictive functionals such as \(f(y \mid x)\), \(S(y \mid x)\), \(Q(\tau \mid x)\), and restricted means.

If parallel_chains = TRUE, chains are run concurrently when the stored MCMC configuration uses more than one chain. If the bundle was built with latent cluster labels monitored, the z_update_every argument controls how frequently those latent indicators are refreshed during sampling.

See also

build_nimble_bundle, mcmc, summary.mixgpd_fit, predict.mixgpd_fit.

Examples

# \donttest{
library(nimble)
y <- abs(rnorm(25)) + 0.1
bundle <- build_nimble_bundle(
  y = y,
  backend = "sb",
  kernel = "normal",
  GPD = FALSE,
  components = 3,
  mcmc = list(niter = 100, nburnin = 50, thin = 1, nchains = 1, seed = 1)
)
fit <- run_mcmc_bundle_manual(bundle, show_progress = FALSE)
#> Defining model
#>   [Note] 'P' is provided in 'constants' but not used in the model code and is being ignored.
#> Building model
#> Setting data and initial values
#> Checking model sizes and dimensions
#>   [Note] This model is not fully initialized. This is not an error.
#>          To see which variables are not initialized, use model$initializeInfo().
#>          For more information on model initialization, see help(modelInitialization).
#> ===== Monitors =====
#> thin = 1: alpha, mean, sd, w
#> ===== Samplers =====
#> conjugate sampler (6)
#>   - mean[]  (3 elements)
#>   - sd[]  (3 elements)
#> categorical sampler (25)
#>   - z[]  (25 elements)
#> RW sampler (3)
#>   - alpha
#>   - v[]  (2 elements)
#> Compiling
#>   [Note] This may take a minute.
#>   [Note] Use 'showCompilerOutput = TRUE' to see C++ compilation details.
#> Compiling
#>   [Note] This may take a minute.
#>   [Note] Use 'showCompilerOutput = TRUE' to see C++ compilation details.
#>   [Warning] To calculate WAIC, set 'WAIC = TRUE', in addition to having enabled WAIC in building the MCMC.
#> running chain 1...
fit
#> MixGPD fit | backend: Stick-Breaking Process | kernel: Normal Distribution | GPD tail: FALSE
#> n = 25 | components = 3 | epsilon = 0.025
#> MCMC: niter=100, nburnin=50, thin=1, nchains=1 
#> Fit
#> Use summary() for posterior summaries; plot() for diagnostics; predict() for predictions.
# }