Skip to contents

Update proposal matrix, start values, and MCMC settings for next stage

Source: R/inference_aux_functions.R
update_inf_stage.Rd

Update proposal matrix, start values, and MCMC settings for next stage

Usage

update_inf_stage(
  results_obj,
  proposal_matrix,
  param_names,
  S_prev = 3000,
  draw_n = 4,
  shrink = 0.8,
  stage,
  n_steps = NULL
)

Arguments

results_obj

Result object returned by inf_run (with MCMC traces)

proposal_matrix

Current proposal variance matrix

param_names

Character vector of parameter names (rownames of proposal_matrix)

S_prev

Integer: number of samples to use for variance-covariance extraction

draw_n

Integer: number of random draws for new start values per chain

shrink

Numeric: shrinkage factor for bounds (e.g. 0.8 retains 80% of interval)

stage

Character: which stage to pull from create_mcmc_params()

n_steps

Integer or NULL: if not NULL, overrides control_params$n_steps

Value

A list with elements:

  • proposal_matrix : updated proposal variance matrix

  • start_values : matrix of new start values (chains x parameters)

  • adaptive_params : list of adaptive settings for this stage

  • control_params : list of control settings for this stage (with n_steps possibly overridden)