The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

Chapter 13: Bayesian inference and decision making with bayestestR

Kjell Nygren

library(glmbayes)
library(bayestestR)

1. Purpose

bayestestR implements a wide range indices for posterior description and decision-oriented summaries (probability of direction, ropes, equivalence tests, etc.). Here we treat the coefficients component of a glmb fit as stored i.i.d. draws and summarize them using bayestestR.

This complements Chapter 12 (optional bayesplot graphics, commented out in the vignette) with tabular summaries you can cite in applied reports.

2. Example: logistic regression posterior for coefficients

data(menarche, package = "MASS")

Age2 <- menarche$Age - 13
Menarche_Model_Data <- data.frame(
  Menarche  = menarche$Menarche,
  Total     = menarche$Total,
  Age2      = Age2
)

ps <- Prior_Setup(
  cbind(Menarche, Total - Menarche) ~ Age2,
  family = binomial(link = "logit"),
  data   = Menarche_Model_Data
)

fit_logit <- glmb(
  cbind(Menarche, Total - Menarche) ~ Age2,
  family  = binomial(link = "logit"),
  pfamily = dNormal(mu = ps$mu, Sigma = ps$Sigma),
  data    = Menarche_Model_Data,
  n             = 800,
  use_parallel  = FALSE
)

coef_draws <- as.data.frame(fit_logit$coefficients)

3. Point and interval summaries

bayestestR::describe_posterior(coef_draws)
#> Summary of Posterior Distribution
#> 
#> Parameter   |    Median |        95% CI |     pd |          ROPE | % in ROPE
#> ----------------------------------------------------------------------------
#> (Intercept) | -7.06e-03 | [-0.13, 0.11] | 54.62% | [-0.10, 0.10] |    95.00%
#> Age2        |      1.62 | [ 1.51, 1.73] |   100% | [-0.10, 0.10] |        0%
bayestestR::hdi(coef_draws, ci = 0.89)
#> Highest Density Interval
#> 
#> Parameter   |       89% HDI
#> ---------------------------
#> (Intercept) | [-0.10, 0.08]
#> Age2        | [ 1.54, 1.72]
bayestestR::rope(coef_draws, range = c(-0.02, 0.02))
#> # Proportion of samples inside the ROPE [-0.02, 0.02]:
#> 
#> Parameter   | Inside ROPE
#> -------------------------
#> (Intercept) |     24.74 %
#> Age2        |      0.00 %
bayestestR::p_direction(coef_draws)
#> Probability of Direction
#> 
#> Parameter   |     pd
#> --------------------
#> (Intercept) | 54.62%
#> Age2        |   100%

Tune rope(..., range = ...) to match substantive “practical equivalence” hypotheses on the logit scale (see also the binomial likelihood discussion in Chapter 09). For richer reporting pipelines (parameters, performance, …), see the easystats ecosystem linking out from ?bayestestR.

See also

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.