precis: Precis of model fit
In rmcelreath/rethinking: Statistical Rethinking book package

precis

R Documentation

Precis of model fit

Description

Displays concise parameter estimate information for an existing model fit.

Usage

precis( model , depth=1 , pars , ci=TRUE , prob=0.89 , 
    corr=FALSE , digits=2 , warn=TRUE )

Arguments

`model`	Fit model object
`depth`	If `1`, suppresses vectors and matrices of parameters. If `2`, displays all parameters
`pars`	Optional character vector of parameter names to display
`ci`	Show quadratic estimate confidence intervals
`prob`	Width of posterior intervals
`corr`	If `TRUE`, show correlations among parameters in output
`digits`	Number of decimal places to display in output
`warn`	If `TRUE`, warns about various things

Details

Creates a table of estimates and standard errors, with optional confidence intervals and parameter correlations. Posterior intervals are quadratic estimates, derived from standard deviations, unless the model uses samples from the posterior distribution, in which case HPDI is used instead.

Can also provide expected value, standard deviation, and HPDI columns for a data frame.

Value

A data frame with a row for each parameter. The n_eff and Rhat columns are calculated by rstan. Rhat4 indicates Rhat as defined in Gelman et al 2013 "Bayesian Data Analysis". This is different from the classic 1992 Gelman and Rubin definition, as Rhat4 uses more information from multiple chains.