cv_quap: Diagnostic trace and rank histogram plots for MCMC output

In rmcelreath/rethinking: Statistical Rethinking book package

Diagnostic trace and rank histogram plots for MCMC output

Description

Cross validation for quap model fits.

Usage

cv_quap( quap_model, lno = 1, pw = FALSE, cores = 1, ... )

Arguments

quap_model	quap model fit
lno	Number of observations to leave out in each fold
pw	Pointwise results (TRUE) or summed across folds (FALSE)
cores	Number of cores to use. If great than 1, uses mclappy with folds
...	Additional arguments to pass to quap

Details

This function constructs cross validation folds from an existing quap model fit and associated data. It then fits the model to each fold and returns either the fit to each fold (when pw=TRUE) or the summed performance across all folds.

The default is leave-one-out cross-validation, when lno=1.

Author(s)

Richard McElreath

rmcelreath/rethinking documentation built on Sept. 18, 2023, 2:01 p.m.