control_multibergm: Set multibergm MCMC options

In brieuclehmann/multibergm: Bayesian Exponential Random Graph Models for Multiple Networks

Set multibergm MCMC options

Description

Auxiliary function used to specify settings controlling multibergm fitting. This allows the user to control various aspects of the fitting algorithm, including MCMC proposal covariance, auxiliary iterations for ERGM simulation, and number of cores to be used in parallel.

Usage

control_multibergm(
  ergm_formula,
  mod_mat,
  constraints = ~.,
  proposal_update_freq = 20,
  proposal_update_max = 1000,
  proposal_rescale = 500,
  init_proposals = NULL,
  aux_iters = 1000,
  n_batches = 1
)

Arguments

ergm_formula	An R formula object, of the form y ~ <model terms>, where y is a network object or a network.list object.
constraints	A one-sided formula specifying one or more constraints on the support of the distribution of the networks being simulated.
proposal_update_freq	How often to update the MCMC proposal covariances.
proposal_update_max	When to stop adaptive the MCMC proposals.
proposal_rescale	When to reset adaptation following an initial burn_in
init_proposals	A list of covariance matrices used to initialise the MCMC proposals.
aux_iters	Number of internal (auxiliary) MCMC iterations used to simulate a network from the model.
n_batches	Number of computing cores available to simulate networks from the model in parallel.
groups	A vector of group memberships

Value

A list containing the following control parameters:

• aux_iters: Number of internal (auxiliary) MCMC iterations used to simulate a network from the model

• proposal: A list of covariance matrices used to generate the MCMC proposals.

• batches: Used to batch parallel runs of ergm simulation

• model: Internal representation of the ergm network model

• clists: ERGM parameters passed to the ergm simulation function

• mh_proposals: ERGM parameters passed to the ergm simulation function

brieuclehmann/multibergm documentation built on June 19, 2024, 6:36 p.m.