loo (version 0.1.0)

loo_and_waic: Approximate LOO-CV and WAIC for Bayesian models

Description

Approximate LOO-CV and WAIC for Bayesian models

Usage

loo_and_waic(log_lik, ...)

Arguments

log_lik

an $S$ by $N$ matrix, where $S$ is the size of the posterior sample (the number of simulations) and $N$ is the number of data points. Typically (but not restricted to be) the object returned by extract_log_lik

...

optional arguments to pass to vgislw. Possible arguments and their defaults are: [object Object],[object Object],[object Object],[object Object],[object Object]

We recommend using the default values for the

Value

• a named list with class 'loo'.

  Returned for both LOO and WAIC are the expected log pointwise predictive density (elpd), the estimated effective number of parameters (p), and the information criteria on the deviance scale (e.g. looic = -2*elpd_loo). Also returned are the pointwise contributions of each of these measures, standard errors, and the estimated shape parameter $k$ for the Pareto fit to the importance ratios for each leave-one-out distribution.

See Also

loo-package, print.loo, loo_and_waic_diff

Examples

log_lik <- extract_log_lik(stanfit)
loo <- loo_and_waic(log_lik)
print(loo, digits = 3)