Skip to contents

Control Parameters for JointODE

Source: R/control.R
JointODE.control.Rd

Construct control parameters for the JointODE optimization algorithm. This function can be called with no arguments to get defaults, or can process a list to fill in missing values with defaults.

Usage

JointODE.control(
  maxit = 200,
  tol = 1e-04,
  verbose = FALSE,
  parallel = FALSE,
  n_cores = 0,
  hazard_quadrature = 1,
  mc_samples = 100,
  mc_burnin = 50,
  .list = NULL,
  ...
)

Arguments

maxit

Maximum number of MCEM iterations (default: 200)

tol

Convergence tolerance. The MCEM algorithm converges when max|theta_new - theta_old| < tol (default: 1e-4)

verbose

Logical or numeric; controls verbosity level. FALSE/0 for silent, TRUE/1 for basic progress, 2 for detailed output (default: FALSE)

parallel

Logical; whether to use parallel computation (default: FALSE)

n_cores

Integer; number of cores to use for parallel computation. If 0, uses all available cores (default: 0)

hazard_quadrature

Integer; number of Simpson sub-intervals per observation interval for hazard integration (default: 1)

mc_samples

Integer; number of MCMC samples per subject for Monte Carlo EM (MCEM). Must be a positive integer. Recommended range: 50-200 for MCEM (default: 100).

mc_burnin

Integer; number of burn-in iterations for the Metropolis-Hastings sampler in the E-step. Discarded before collecting mc_samples draws (default: 50).

.list

Optional list of control parameters to process

...

Additional control parameters

Value

A list of control parameters with all defaults filled in

See also

JointODE

Examples

# Default settings
control <- JointODE.control()

# Custom settings for faster exploration
control <- JointODE.control(maxit = 30, tol = 1e-4)

# Verbose output for debugging
control <- JointODE.control(verbose = TRUE)

# Parallel computation
control <- JointODE.control(parallel = TRUE, n_cores = 4)

# Process an existing list
my_list <- list(maxit = 200)
control <- JointODE.control(.list = my_list)