Skip to contents

stan_variational

stan_variational.Rd

Estimate parameters using Stan's variational inference algorithms

Usage

stan_variational(
  fn,
  par_inits = NULL,
  n_pars = NULL,
  additional_args = list(),
  algorithm = "meanfield",
  grad_fun = NULL,
  lower = -Inf,
  upper = Inf,
  eval_standalone = FALSE,
  globals = TRUE,
  packages = NULL,
  seed = NULL,
  refresh = NULL,
  quiet = FALSE,
  output_dir = NULL,
  output_basename = NULL,
  sig_figs = NULL,
  iter = NULL,
  grad_samples = NULL,
  elbo_samples = NULL,
  eta = NULL,
  adapt_engaged = NULL,
  adapt_iter = NULL,
  tol_rel_obj = NULL,
  eval_elbo = NULL,
  output_samples = NULL
)

Arguments

fn

Function to estimate parameters for

par_inits

Initial values for parameters. This can either be a numeric vector of initial values (which will be used for all chains), a list of numeric vectors (of length equal to the number of chains), a function taking a single argument (the chain ID) and returning a numeric vector of initial values, or NULL (in which case Stan will generate initial values automatically). (must be specified if n_pars is NULL)

n_pars

Number of parameters to estimate (must be specified if par_inits is NULL)

additional_args

List of additional arguments to pass to the function

algorithm

(string) The variational inference algorithm. One of "meanfield" or "fullrank".

grad_fun

Function calculating gradients w.r.t. each parameter

lower

Lower bound constraint(s) for parameters

upper

Upper bound constraint(s) for parameters

eval_standalone

(logical) Whether to evaluate the function in a separate R session. Defaults to FALSE.

globals

(optional) a logical, a character vector, or a named list to control how globals are handled when evaluating functions in a separate R session. Ignored if eval_standalone = FALSE. For details, see section 'Globals used by future expressions' in the help for future::future().

packages

(optional) a character vector specifying packages to be attached in the R environment evaluating the function. Ignored if eval_standalone = FALSE.

seed

Random seed

refresh

Number of iterations for printing

quiet

(logical) Whether to suppress Stan's output

output_dir

Directory to store outputs

output_basename

Basename to use for output files

sig_figs

Number of significant digits to use for printing

iter

(positive integer) The maximum number of iterations.

grad_samples

(positive integer) The number of samples for Monte Carlo estimate of gradients.

elbo_samples

(positive integer) The number of samples for Monte Carlo estimate of ELBO (objective function).

eta

(positive real) The step size weighting parameter for adaptive step size sequence.

adapt_engaged

(logical) Do warmup adaptation?

adapt_iter

(positive integer) The maximum number of adaptation iterations.

tol_rel_obj

(positive real) Convergence tolerance on the relative norm of the objective.

eval_elbo

(positive integer) Evaluate ELBO every Nth iteration.

output_samples

(positive integer) Number of approximate posterior samples to draw and save.

Value

StanVariational object