Summaries and intervals of draws

posterior_summary() reduces draws to estimate, error and quantiles in brms's column layout (Estimate, Est.Error, Q2.5, Q97.5). A matrix has variables in columns and draws in rows and gives one row per variable. A three-dimensional array has draws in its first margin, as ranef(summary = FALSE) returns them, and gives a variables x statistics x third-margin array, as in brms.

Usage

posterior_summary(x, ...)

# Default S3 method
posterior_summary(x, probs = c(0.025, 0.975), robust = FALSE, ...)

# S3 method for class 'frmtmb_fit'
posterior_summary(x, ...)

# S3 method for class 'frmtmb_multiple'
posterior_summary(x, ...)

Arguments

x

A matrix or three-dimensional array of draws.

...

Refused: an argument the method does not have is an error naming it, rather than silently changing nothing.

probs

Quantiles to report.

robust

If TRUE, median and MAD instead of mean and SD.

Value

A matrix with one row per variable, or an array, as above.

Details

The computation is brms's posterior_summary.default(), statistic for statistic: mean and sd, or median and mad when robust, then quantile, each with na.rm = TRUE.

The method for posterior draws of a fitted model is in the frmtmb.sample package, along with the sampler that produces them. A maximum-likelihood fit has no draws, and its method says so.

Examples

# any matrix of draws: rows are draws, columns are variables
m <- cbind(a = rnorm(500), b = rnorm(500, 2))
posterior_summary(m)
#>     Estimate Est.Error       Q2.5    Q97.5
#> a 0.01077349  1.000574 -1.9062756 2.093029
#> b 1.94079266  0.979340  0.0894709 3.910332
posterior_summary(m, robust = TRUE)
#>      Estimate Est.Error       Q2.5    Q97.5
#> a -0.01590824 0.9905923 -1.9062756 2.093029
#> b  1.91763189 1.0087451  0.0894709 3.910332

# a maximum-likelihood fit has no draws to summarize
dd <- data.frame(y = rnorm(40), x = rnorm(40))
try(posterior_summary(frm(bf(y ~ x) + gaussian(), data = dd)))
#> Error : posterior_summary() needs posterior draws and a frmtmb_fit has none: frm() is maximum likelihood, so it carries one parameter vector rather than chains. Install frmtmb.sample and sample first, with posterior_summary(frmtmb.sample::frm_sample(fit)); for the point estimates and their covariance use fixef(), vcov() or confint() on the fit itself