calcPercChangeMCMC: calcPercChangeMCMC

In SOLV-Code/MetricsTest: Functions for WSP Metrics

Description

MCMC version of trend metric calculation (replaces old version based on PBSModelling and Metropolis algorithm)

Usage

calcPercChangeMCMC(
  vec.in,
  model.in = NULL,
  perc.change.bm = -25,
  na.skip = FALSE,
  out.type = "short",
  mcmc.plots = FALSE,
  convergence.check = FALSE
)

Arguments

vec.in	vector with numeric values
model.in	if NULL, use the BUGS code in the built in function trend.bugs.1()
perc.change.bm	benchmark for Prob(Decl>BM), default = -25
na.skip	if TRUE, skip the calculations when vec.in contains any NA
out.type	"short" or "long". "short" gives summary table of posterior plus "PercChange" and "ProbDecl" "long" also includes the full mcmc samples and the jags output object
mcmc.plots	if true, create the standard series of MCMC diagnostic plots Note that these are printed to the default device (i.e. need to external wrap the function call inside a pdf /dev.off call) To get a plot of the model fit, run this function with out.type = "long", and then use plot.trend.fit().
convergence.check	if TRUE, do an automated convergence check

SOLV-Code/MetricsTest documentation built on Feb. 19, 2021, 10:12 p.m.