estimate predictive error

The loo package (R) provides efficient approximate leave-one-out cross-validation (LOO), approximate standard errors for estimated predictive errors (including comparisons), and the widely applicable information criterion (WAIC).

The loo package can be used to estimate predictive error of any MCMC item-level log likelihood output (including but not limited to Stan, JAGS, BUGS, MCMCPack, NIMBLE, emcee, and SAS).


Visit mc-stan.org/loo for vignettes/tutorials, function documentation, and other information about the package.

Download and Get Started

The loo package, documentation, and examples are available from CRAN.

Source Code and Issue Tracker

The loo package’s source code and issue tracker are hosted by GitHub.


The loo package is open-source licensed under the