Pointwise log-likelihood

log_lik() is the log-likelihood of every observation at every posterior draw, the matrix loo() and waic() are computed from. A frmtmb_fit is one maximum-likelihood parameter vector rather than a posterior, so this refuses and names the route to draws, as loo() and waic() do; the estimator itself is in the frmtmb.sample package, on this same generic.

Usage

log_lik(object, ...)

# S3 method for class 'frmtmb_fit'
log_lik(object, ...)

Arguments

object

A frmtmb_fit, or (with frmtmb.sample loaded) draws.

...

Passed to methods.

Value

This method signals an error on a maximum-likelihood fit.

Details

The generic lives here rather than in frmtmb.sample so that a script ported from brms is told to sample before it has sampled. Without it log_lik(fit) stopped at "could not find function", which names neither the reason nor the route.

See also

loo() and waic(), which average this matrix over draws; logLik() for the maximum-likelihood total.

Examples

set.seed(1)
dd <- data.frame(x = rnorm(40))
dd$y <- rnorm(40, 1 + 0.5 * dd$x, 1)
fit <- frm(bf(y ~ x) + gaussian(), data = dd)

# the maximum-likelihood total is available directly
logLik(fit)
#> 'log Lik.' -52.08647 (df=3)
# the pointwise posterior matrix needs draws, and says so
try(log_lik(fit))
#> Error : log_lik() is the pointwise likelihood at every posterior draw, and this is a maximum-likelihood fit with one parameter vector. Sample first, with frmtmb.sample::log_lik(frmtmb.sample::frm_sample(fit)) once that package is installed; logLik() is the maximum-likelihood total already on the fit