ggdmc (version 0.2.5.8)

gelman: Potential scale reduction factor

Description

gelman function calls the function, gelman.diag in the coda package to calculates PSRF.

Usage

gelman(x, hyper = FALSE, start = 1, end = NA, confidence = 0.95,
  transform = TRUE, autoburnin = FALSE, multivariate = TRUE,
  split = TRUE, subchain = FALSE, nsubchain = 3, digits = 2,
  verbose = FALSE, ...)
hgelman(x, start = 1, end = NA, confidence = 0.95,
  transform = TRUE, autoburnin = FALSE, split = TRUE,
  subchain = FALSE, nsubchain = 3, digits = 2, verbose = FALSE,
  ...)

Arguments

x

posterior samples

hyper

a Boolean switch, indicating posterior samples are from hierarchical modeling

start

start iteration

end

end iteration

confidence

confident inteval

transform

turn on transform

autoburnin

turn on auto burnin

multivariate

multivariate Boolean switch

split

split whether split mcmc chains; When split is TRUE, the function doubles the number of chains by spliting into 1st and 2nd halves.

subchain

whether only calculate a subset of chains

nsubchain

indicate how many chains in a subset

digits

print out how many digits

verbose

print more information

...

arguments passing to coda gelman.diag.

Examples

# NOT RUN {
rhat1 <- hgelman(hsam); rhat1
rhat2 <- hgelman(hsam, end = 51); rhat2
rhat3 <- hgelman(hsam, confidence = .90); rhat3
rhat4 <- hgelman(hsam, transform = FALSE); rhat4
rhat5 <- hgelman(hsam, autoburnin = TRUE); rhat5
rhat6 <- hgelman(hsam, split = FALSE); rhat6
rhat7 <- hgelman(hsam, subchain = TRUE); rhat7
rhat8 <- hgelman(hsam, subchain = TRUE, nsubchain = 4);
rhat9 <- hgelman(hsam, subchain = TRUE, nsubchain = 4,
digits = 1, verbose = TRUE);

hat1 <- gelman(hsam[[1]], multivariate = FALSE); hat1
hat2 <- gelman(hsam[[1]], hyper = TRUE, verbose = TRUE); hat2
hat3 <- gelman(hsam, hyper = TRUE, verbose = TRUE); hat3
hat4 <- gelman(hsam, multivariate = TRUE, verbose = FALSE);
hat5 <- gelman(hsam, multivariate = FALSE, verbose = FALSE);
hat6 <- gelman(hsam, multivariate = FALSE, verbose = TRUE);
hat7 <- gelman(hsam, multivariate = T, verbose = TRUE);
# }