Widely Applicable Information Criterion (WAIC)

Compute the widely applicable information criterion (WAIC) based on the posterior likelihood using the loo package. For more details see waic.

# S3 method for brmsfit
waic(
  x,
  ...,
  compare = TRUE,
  resp = NULL,
  pointwise = FALSE,
  model_names = NULL
)

Arguments

x: A brmsfit object.
...: More brmsfit objects or further arguments passed to the underlying post-processing functions. In particular, see prepare_predictions for further supported arguments.
compare: A flag indicating if the information criteria of the models should be compared to each other via loo_compare.
resp: Optional names of response variables. If specified, predictions are performed only for the specified response variables.
pointwise: A flag indicating whether to compute the full log-likelihood matrix at once or separately for each observation. The latter approach is usually considerably slower but requires much less working memory. Accordingly, if one runs into memory issues, pointwise = TRUE is the way to go.
model_names: If NULL (the default) will use model names derived from deparsing the call. Otherwise will use the passed values as model names.

Value

If just one object is provided, an object of class loo. If multiple objects are provided, an object of class loolist.

Details

See loo_compare for details on model comparisons. For brmsfit objects, WAIC is an alias of waic. Use method add_criterion to store information criteria in the fitted model object for later usage.

References

Vehtari, A., Gelman, A., & Gabry J. (2016). Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. In Statistics and Computing, doi:10.1007/s11222-016-9696-4. arXiv preprint arXiv:1507.04544.

Gelman, A., Hwang, J., & Vehtari, A. (2014). Understanding predictive information criteria for Bayesian models. Statistics and Computing, 24, 997-1016.

Watanabe, S. (2010). Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular learning theory. The Journal of Machine Learning Research, 11, 3571-3594.

Examples

if (FALSE) {
# model with population-level effects only
fit1 <- brm(rating ~ treat + period + carry,
            data = inhaler)
(waic1 <- waic(fit1))

# model with an additional varying intercept for subjects
fit2 <- brm(rating ~ treat + period + carry + (1|subject),
            data = inhaler)
(waic2 <- waic(fit2))

# compare both models
loo_compare(waic1, waic2)
}