Add leave-one-out (LOO; Vehtari et al., 2017) and/or widely applicable information criteria (WAIC; Watanabe, 2010) to fitted model objects.
Usage
add_criterion(
x,
criterion = c("loo", "waic"),
overwrite = FALSE,
save = TRUE,
...,
r_eff = NA
)
Arguments
- x
A measrfit object.
- criterion
A vector of criteria to calculate and add to the model object.
- overwrite
Logical. Indicates whether existing criteria should be overwritten. For example, if LOO has already been added to the model object
add_criterion()
is called again withcriterion = "loo"
, should LOO be recalculated? Default isFALSE
.- save
Logical. Only relevant if a file was specified in the measrfit object passed to
x
. IfTRUE
(the default), the model is re-saved to the specified file when new criteria are added to the R object. IfFALSE
, the new criteria will be added to the R object, but the saved file will not be updated.- ...
Additional arguments passed to
loo::loo.array()
orloo::waic.array()
.- r_eff
Vector of relative effective sample size estimates for the likelihood (
exp(log_lik)
) of each observation. This is related to the relative efficiency of estimating the normalizing term in self-normalizing importance sampling when using posterior draws obtained with MCMC. If MCMC draws are used andr_eff
is not provided then the reported PSIS effective sample sizes and Monte Carlo error estimates will be over-optimistic. If the posterior draws are independent thenr_eff=1
and can be omitted. The warning message thrown whenr_eff
is not specified can be disabled by settingr_eff
toNA
. See therelative_eff()
helper functions for computingr_eff
.
Value
A modified measrfit object with the criteria
slot populated with
the specified criteria.
References
Vehtari, A., Gelman, A., & Gabry, J. (2017). Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Statistics and Computing, 27(5), 1413-1432. https://doi.org/10.1007/s11222-016-9696-4
Watanabe, S. (2010). Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular learning theory. Journal of Machine Learning Research, 11(116), 3571-3594. http://jmlr.org/papers/v11/watanabe10a.html