loo
Efficient Leave-One-Out Cross-Validation and WAIC for Bayesian Models
Efficient approximate leave-one-out cross-validation (LOO) for Bayesian models fit using Markov chain Monte Carlo, as described in Vehtari, Gelman, and Gabry (2017) doi:10.1007/s11222-016-9696-4. The approximation uses Pareto smoothed importance sampling (PSIS), a new procedure for regularizing importance weights. As a byproduct of the calculations, we also obtain approximate standard errors for estimated predictive errors and for the comparison of predictive errors between models. The package also provides methods for using stacking and other model weighting techniques to average Bayesian predictive distributions.
- Version2.8.0
- R version≥ 3.1.2
- LicenseGPL (≥ 3)
- Needs compilation?No
- loo citation info
- Last release07/03/2024
Documentation
- VignetteHoldout validation and K-fold cross-validation of Stan programs with the loo package
- VignetteUsing the loo package
- VignetteUsing Leave-one-out cross-validation for large data
- VignetteApproximate leave-future-out cross-validation for Bayesian time series models
- VignetteMixture IS leave-one-out cross-validation for high-dimensional Bayesian models
- VignetteAvoiding model refits in leave-one-out cross-validation with moment matching
- VignetteLeave-one-out cross-validation for non-factorized models
- VignetteBayesian Stacking and Pseudo-BMA weights
- VignetteWriting Stan programs for use with the loo package
- MaterialNEWS
- In ViewsBayesian
Team
Jonah Gabry
MaintainerShow author detailsBen Goodrich
Show author detailsRolesContributorPaul-Christian Bürkner
Show author detailsRolesAuthorJuho Piironen
Show author detailsRolesContributorMåns Magnusson
Show author detailsRolesAuthorAndrew Gelman
Show author detailsRolesAuthorAki Vehtari
Show author detailsRolesAuthorYuling Yao
Show author detailsRolesAuthorTopi Paananen
Show author detailsRolesAuthorBruno Nicenboim
Show author detailsRolesContributorLeevi Lindgren
Show author detailsRolesContributor
Insights
Last 30 days
This package has been downloaded 52,839 times in the last 30 days. This work is reaching a lot of screens. A significant achievement indeed! The following heatmap shows the distribution of downloads per day. Yesterday, it was downloaded 2,211 times.
The following line graph shows the downloads per day. You can hover over the graph to see the exact number of downloads per day.
Last 365 days
This package has been downloaded 571,104 times in the last 365 days. This is starting to look legendary. Enough downloads to inspire a dedicated panel discussion! The day with the most downloads was Mar 25, 2025 with 2,744 downloads.
The following line graph shows the downloads per day. You can hover over the graph to see the exact number of downloads per day.
Data provided by CRAN
Binaries
Dependencies
- Imports3 packages
- Suggests9 packages
- Reverse Depends4 packages
- Reverse Imports50 packages
- Reverse Suggests11 packages