Skip to contents

Run MCMC sampling on a compiled DPMirt model

dpmirt_sample.Rd

Executes MCMC sampling using the compiled model. This is the lightweight step that can be called repeatedly from the same compiled model (compile-once, sample-many pattern).

Usage

dpmirt_sample(
  compiled,
  niter = 10000L,
  nburnin = 2000L,
  thin = 1L,
  thin2 = NULL,
  seed = NULL,
  reset = TRUE,
  verbose = TRUE,
  inits = NULL,
  init_seed = NULL,
  init_strategy = NULL,
  ...
)

# S3 method for class 'dpmirt_samples'
print(x, ...)

Arguments

compiled

A dpmirt_compiled object from dpmirt_compile.

niter

Integer. Total number of MCMC iterations.

nburnin

Integer. Number of burn-in iterations to discard.

thin

Integer. Thinning interval for main monitors.

thin2

Integer or NULL. Thinning interval for monitors2 (eta/theta). If NULL, uses the same value as thin.

seed

Integer or NULL. Random seed for reproducibility.

reset

Logical. If TRUE (default), reset sample and WAIC storage before sampling. Set to FALSE to continue from the compiled object's current sampler state and append to existing sample storage.

verbose

Logical. Print progress messages.

inits

Optional list of initial values to apply before sampling. This is mainly used internally for chain-specific starts and requires reset = TRUE.

init_seed

Optional integer seed used to generate inits, stored as provenance.

init_strategy

Optional character label describing how initial values were chosen, stored as provenance.

...

Additional arguments (currently unused).

x

A dpmirt_samples object.

Value

A dpmirt_samples S3 object containing:

samples

Matrix of posterior samples from main monitors.

samples2

Matrix of posterior samples from thinned monitors (eta).

waic

WAIC value if computed, otherwise NULL.

sampling_time

Time taken for sampling.

mcmc_control

List of MCMC settings used.

model_config

Reference to model configuration.

compiled

Reference to compiled object (for resume).

schema_version

Draw-storage schema version.

chain_info

Data frame describing retained rows by chain.

draw_index

Main/theta row maps with chain and MCMC iteration IDs.

run_history

Data frame describing sampling runs.

See also

dpmirt_compile, dpmirt_resume, dpmirt_rescale

Other model fitting: dpmirt(), dpmirt_compile(), dpmirt_fit-methods, dpmirt_resume(), dpmirt_spec()

Examples

if (FALSE) { # \dontrun{
sim <- dpmirt_simulate(200, 20, model = "rasch", seed = 42)
spec <- dpmirt_spec(sim$response, model = "rasch", prior = "normal")
compiled <- dpmirt_compile(spec)

# Run MCMC sampling
samples <- dpmirt_sample(compiled, niter = 5000, nburnin = 1000, seed = 123)
print(samples)
} # }