run_until_converged: Run a sampler until it has converged, or until the maximum...

In ashiklom/mvtraits: Fit multivariate models to partially missing trait data

Description

Run a sampler until it has converged, or until the maximum number of attempts has been exceeded

Usage

run_until_converged(sampler, model_type, inits, nchains, max_attempts,
  save_progress, threshold, keep_samples, autofit)

Arguments

sampler	A named list containing fun (the function to call as the sampler), init_fun (the function used to generate the initial conditions), and args (a named list of arguments for the sampler function).
model_type	Type of model. Must be either "multi" or "hier".
inits	Named list of initial conditions.
nchains	Number of chains (default = 3)
max_attempts	Maximum number of attempts for autofit (default = 10)
save_progress	If not NULL, path to file in which to save current
threshold	Gelman PSRF threshold for convergence (default = 1.15)
keep_samples	Maximum number of samples to keep in results (default = Inf)
autofit	(Logical) If TRUE, keep re-trying until model has converged (default = FALSE)

Value

List of sample chains. Each sample chain is a named list of sampled parameters.

Author(s)

Alexey Shiklomanov

ashiklom/mvtraits documentation built on Sept. 4, 2019, 7:43 a.m.