Some diagnostics for fitted dose-response models

Summary method for objects of class gvs, as returned by gibbs. Produces a text-based summary of: (1) effective sample sizes, (2) acceptance rates, (3) model convergence, and (4) posterior model probabilities. Note that gibbs does not implement covariate selection. As a result, posterior inclusion probabilities (PIPs) are not returned here, contrary to summary.rjtrace.

# S3 method for gvs
summary(
  gvs.obj,
  eff.n = TRUE,
  accept.rate = TRUE,
  convergence = TRUE,
  gelman.rubin = 1.1,
  model.ranks = TRUE,
  n.top = NULL
)

Arguments

gvs.obj	Input trace object, as returned by gibbs.
eff.n	Logical. If TRUE, returns estimates of the effective sample size for each model parameter.
accept.rate	Logical. If TRUE, returns the acceptance rate (calculated after burn-in) for each model parameter.
convergence	Logical. If TRUE, assesses convergence using the multivariate potential scale reduction factor (Gelman-Rubin statistic), as implemented in gelman.diag.
gelman.rubin	Threshold for determining convergence based on the Gelman-Rubin statistic. Defaults to 1.1.
model.ranks	Logical. If TRUE, returns a summary of posterior model probabilities and associated model rankings.
n.top	Number of top-ranking models to display when model.ranks = TRUE.

Value

A detailed summary, printed to the R console.

See also

simulate_data example_brs summary.brsdata

Author

Phil J. Bouchet

Examples

if (FALSE) {
library(espresso)

# Simulate data for two species
mydat <- simulate_data(n.species = 2,
                       n.whales = 16,
                       min.trials = 1,
                       max.trials = 3,
                       covariates = list(exposed = c(0, 5), range = 0.5),
                       mu = c(101, 158),
                       phi = 20,
                       sigma = 20,
                       Lc = c(60, 65),
                       Rc = c(210, 211),
                       seed = 58697)
summary(mydat)

# Model selection by GVS
gvs.model <- gibbs(dat = mydat,
             random.effects = FALSE,
             include.covariates = FALSE,
             mcmc.n = 1000,
             burnin = 500)

summary(gvs.model)
}