sma_assess: Calculates posterior predictive checks (Bayesian p-values).

In audrey-b/simanalyse: Facilitate Model Comparisons and Simulation Studies

Calculates posterior predictive checks (Bayesian p-values).

Description

This function is in development. Use at your own risk.

Usage

sma_assess(
  object,
  expr,
  data,
  code,
  constants = nlist(),
  nsamples = 100L,
  statistic = "FT"
)

Arguments

object	An mcmcr object of results
expr	An expression to calculate the expectations, e.g. "X = mu" if the expectation of X is mu
data	A dataset
code	Code to simulate data
constants	Constants to pass with the code
nsamples	Number of samples to take from the posterior distribution
statistic	A string, one of "FT" (Freeman-Tukey), "LR" (Likelihood Ratio), or "chi2" (Chi-square)

Value

A flag.

Examples

code="for(i in 1:10){
a[i] ~ dpois(mu)}"
sims <- sims::sims_simulate(code, parameters=nlist(mu=5), nsims=1)
res <- sma_analyse(sims, 
                           code, 
                           "mu ~ dunif(0,10)", 
                           monitor = "mu",
                           deviance=FALSE,
                           mode=sma_set_mode("quick", n.save=300, max.iter=300))
sma_assess(res, expr="a = rep(mu, 10)", sims, code, nsamples=10)

audrey-b/simanalyse documentation built on May 20, 2022, 7:45 p.m.