Raw File
Tip revision: e1fa15d202de277bb07e58bb3013557724072b2b authored by Dominique Makowski on 22 September 2019, 15:30:05 UTC
version 0.3.0
Tip revision: e1fa15d
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/diagnostic_posterior.R
\title{Posteriors Sampling Diagnostic}
diagnostic_posterior(posteriors, diagnostic = c("ESS", "Rhat"), ...)

\method{diagnostic_posterior}{stanreg}(posteriors, diagnostic = "all",
  effects = c("fixed", "random", "all"), parameters = NULL, ...)

\method{diagnostic_posterior}{brmsfit}(posteriors, diagnostic = "all",
  effects = c("fixed", "random", "all"), component = c("conditional",
  "zi", "zero_inflated", "all"), parameters = NULL, ...)
\item{posteriors}{A stanreg or brms model.}

\item{diagnostic}{Diagnostic metrics to compute.  Character (vector) or list with one or more of these options: \code{"ESS"}, \code{"Rhat"}, \code{"MCSE"} or \code{"all"}.}

\item{...}{Currently not used.}

\item{effects}{Should parameters for fixed effects, random effects
or both be returned? Only applies to mixed models. May be abbreviated.}

\item{parameters}{Regular expression pattern that describes the parameters that
should be returned.}

\item{component}{Should all parameters, parameters for the
conditional model, the zero-inflated part of the model, the dispersion
term or the instrumental variables be returned? Applies to models
with zero-inflated and/or dispersion formula, or to models with instrumental
variable (so called fixed-effects regressions). May be abbreviated.}
Extract diagnostic metrics (Effective Sample Size (\code{ESS}), \code{Rhat} and Monte Carlo Standard Error \code{MCSE}).
\strong{Effective Sample (ESS)} should be as large as possible, altough for most applications, an effective sample size greater than 1000 is sufficient for stable estimates (Bürkner, 2017). The ESS corresponds to the number of independent samples with the same estimation power as the N autocorrelated samples. It is is a measure of \dQuote{how much independent information there is in autocorrelated chains} (\emph{Kruschke 2015, p182-3}).
  \cr \cr
  \strong{Rhat} should be the closest to 1. It should not be larger than 1.1 (Gelman and Rubin, 1992) or 1.01 (Vehtari et al., 2019). The split R-hat statistic quantifies the consistency of an ensemble of Markov chains.
  \cr \cr
  \strong{Monte Carlo Standard Error (MCSE)} is another measure of accuracy of the chains. It is defined as standard deviation of the chains divided by their effective sample size (the formula for \code{mcse()} is from Kruschke 2015, p. 187). The MCSE \dQuote{provides a quantitative suggestion of how big the estimation noise is}.
# rstanarm models
# -----------------------------------------------
model <- stan_glm(mpg ~ wt + gear, data = mtcars, chains = 2, iter = 200, refresh = 0)
# brms models
# -----------------------------------------------
model <- brms::brm(mpg ~ wt + cyl, data = mtcars)

  \item Gelman, A., & Rubin, D. B. (1992). Inference from iterative simulation using multiple sequences. Statistical science, 7(4), 457-472.
  \item Vehtari, A., Gelman, A., Simpson, D., Carpenter, B., \& Bürkner, P. C. (2019). Rank-normalization, folding, and localization: An improved Rhat for assessing convergence of MCMC. arXiv preprint arXiv:1903.08008.
  \item Kruschke, J. (2014). Doing Bayesian data analysis: A tutorial with R, JAGS, and Stan. Academic Press.
back to top