ipd.run.parallel: Run the model using the ipd object with parallel computation

In bipd: Bayesian Individual Patient Data Meta-Analysis using 'JAGS'

Run the model using the ipd object with parallel computation

Description

This function runs the model through parallel computation using dclone R package. Before running this function, we need to specify data, prior, JAGS code, etc. using ipd.model type function.

Usage

ipd.run.parallel(
  ipd,
  pars.save = NULL,
  inits = NULL,
  n.chains = 2,
  n.adapt = 1000,
  n.burnin = 1000,
  n.iter = 10000
)

Arguments

ipd	ipd object created from ipd.model type function
pars.save	parameters to save. For instance, "beta" - coefficients for main effects; "gamma" - coefficients for effect modifiers; "delta" - average treatment effect
inits	initial values specified for the parameters to save
n.chains	number of MCMC chains to sample
n.adapt	number of iterations for adaptation (Note that the samples from adaptation phase is non-Markovian and do not constitute a Markov chain)
n.burnin	number of iterations for burn-in
n.iter	number of iterations to run after the adaptation

Value

MCMC samples stored using JAGS. The returned samples have the form of mcmc.list and coda functions can be directly applied.

Examples

ds <- generate_ipdma_example(type = "continuous")
ipd <- with(ds, ipdma.model.onestage(y = y, study = studyid, treat = treat, X = cbind(z1, z2), 
response = "normal", shrinkage = "none"))

samples <- ipd.run.parallel(ipd, n.chains = 2, n.burnin = 500, n.iter = 5000)

bipd documentation built on June 7, 2022, 1:11 a.m.