par.trace.samples: par.trace.samples

In dlandy/demographicModeling: Supports Bayesian modeling of Qualtrics demographic data

Description

Composite function that handles creating, initing, and running jags models in parallel, then collects the resulting samples.

Usage

par.trace.samples(file, data, n.adapt = 500, inits = NULL, n.iter = 1000,
  n.samples = 500, thin = 1, monitor, n.breaks = c(10, 2, 5),
  predictionVariable = FALSE)

Arguments

file	A jags file
data	Any data values, in a list, as though you were passing them to jags.model (you are)
n.adapt	The number of adapts to do
inits	Initial parameters
n.iter	A number of iterations to run on the update cycle
n.samples	The number of samples to collect.
thin	How much to thin
monitor	What variables do we want to monitor?
n.breaks	How often should we pause to update the user about progress?
predictionVariable	If there is a prediction variable, that variable can be produced in a special cycle, to avoid generating too much data

Value

a list of samples

Examples

results <- par.trace.samples(file=paste(workingDir, jagsFile , sep="")
    , data=jagsData
    , monitor=monitor
    , n.iter = updateNum
    , n.adapt = n.adapt
    , n.samples = nSamplesCoda
    , n.breaks=5
    , thin=thinning
    )

dlandy/demographicModeling documentation built on May 16, 2019, 7:10 p.m.