Calculates central probability intervals for each parameter for each stanfit object

The function is a trivial wrapper around rstantools::posterior_interval() for a list of stanfit objects.

Usage

posterior_intervals(stanfit_list, prob, params_desired = NA, ...)

Arguments

stanfit_list: A list of stanfit objects e.g. as returned by rstan::sampling(). The function's name reflects the assumption that these contain samples from the posterior, but they could contain samples from the prior.
prob: A single number between 0 and 1 specifying the amount of probability to be contained in the interval. e.g. Specifying 0.95 results in an interval that spans the 2.5th to 97.5th percentiles of the distribution.
params_desired: A character vector with the names of those parameters in the stanfit object that should be included in the output. The default value of NA means all parameters are included except "lp__" (which is not actually a parameter: Stan uses it to record the value of the target probability density).
...: Additional arguments to pass to rstantools::posterior_interval().

Value

A 3D array: the first index for the parameter, the second for the percentile, the third for the stanfit object.

Examples

   eg <- mastiff::stan_example_regression
   posterior_intervals(list(eg$posterior_samples, eg$prior_samples), prob =
   0.95)
#> , , 1
#> 
#>            2.5%    97.5%
#> m     2.9321956 3.084793
#> c     1.1434339 2.989257
#> sigma 0.7249795 1.425864
#> 
#> , , 2
#> 
#>             2.5%    97.5%
#> m     0.18343293 5.820427
#> c     0.09083891 3.891102
#> sigma 0.05382647 1.959117
#>