read.mcmc: Read in data from a set of MCMC runs

In mcgibbsit: Warnes and Raftery's 'MCGibbsit' MCMC Run Length and Convergence Diagnostic

Read in data from a set of MCMC runs

Description

Read in data from a set of MCMC runs and create an mcmc.list object.

Usage

read.mcmc(
  nc,
  sourcepattern,
  ...,
  col.names,
  start = 1,
  end = nrow(tmp)/numComponents * thin,
  thin = 1,
  numComponents = 1
)

Arguments

nc	Number of MCMC sampler files to read
sourcepattern	MCMC data file name pattern.
...	Arguments to be passed to read.table when loading MCMC sampler data.
col.names	Data file column names (optional)
start, end, thin	See documentation for mcmc
numComponents	Number of component samplers.

Details

This function reads in the states output from one or more MCMC samplers and creates a single mcmc.list object. sourcepattern will be used as a filename pattern with # replaced by the sampler number. EG, sourcepattern="MCMC.#.csv" will be converted to "MCMC.1.csv", "MCMC.2.csv", etc.

The function read.table is used to read in the data. Options for read.table may be included as part of the call to read.mcmc.

The start, end, and thin arguments can be used to annotate the MCMC samplers with additional information.

Value

An mcmc.list object containing nc component mcmc objects.

Author(s)

Gregory R. Warnes [e-mail]

See Also

mcmc, mcmc.list, read.table

Examples

###
# Create example data files for 20 independent chains
# with serial correlation of 0.
###

set.seed(42)
tmpdir <- tempdir()

for(i in 1:20){
  x <- matrix(rnorm(1000), ncol=4)
  
  x[,4] <- x[,4] + 1/3 * (x[,1] + x[,2] + x[,3])
  
  colnames(x) <- c("alpha","beta","gamma", "nu")
  
  write.table(
    x,
    file = file.path(
      tmpdir,
      paste("mcmc", i, "csv", sep=".")
      ),
    sep = ",",
    row.names=FALSE
  )
}

# Read them back in as an mcmc.list object
data <- read.mcmc(20, file.path(tmpdir, "mcmc.#.csv"), sep=",")

summary(data)

mcgibbsit documentation built on Sept. 25, 2023, 5:06 p.m.