WAIC function

WAIC and LOO

The function computes widely applicable information criterion (WAIC) and efficient approximate leave-one-out cross-validation (LOO) from fitted regression model objects of class flexreg.

WAIC(model, ...)

## S3 method for class 'WAIC.flexreg'
print(x, ...)

Arguments

• model: an object (or a list of objects) of class flexreg, usually the result of flexreg or flexreg_binom functions.
• ...: additional arguments.
• x: an object of class WAIC.flexreg, usually the result of WAIC.

Returns

A named list with components from loo and waic.

Details

This function takes advantage of the loo package to compute the widely applicable information criterion (WAIC) and leave-one-out cross-validation (LOO) for objects of class flexreg. If a list of two or more objects of class flexreg is provided, the function returns the difference in their expected predictive accuracy (see loo_compare for further details).

Examples

## Not run:

data("Reading")
FB <- flexreg(accuracy.adj ~ iq, data = Reading, type="FB", n.iter=1000)
WAIC(FB)
## End(Not run)

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. doi:10.1007/s11222-016-9696-4