loo_cv: Approximate leave-one-out cross-validation (LOO) using...

In bentrueman/bgamcar1: Fit Bayesian GAMs With CAR(1) Errors To Censored Data

Approximate leave-one-out cross-validation (LOO) using loo::loo().

Description

Approximate leave-one-out cross-validation (LOO) using loo::loo().

Usage

loo_cv(input, object, censoring = TRUE, ...)

Arguments

input	The dataframe of observations used to generate the model.
object	A brms model object.
censoring	Logical. Is the response left-censored?
...	Arguments passed on to add_pred_draws_car1().

Value

An object of class loo.

Examples

library("brms")
seed <- 1
data <- read.csv(paste0(system.file("extdata", package = "bgamcar1"), "/data.csv"))
fit <- fit_stan_model(
   paste0(system.file("extdata", package = "bgamcar1"), "/test"),
   seed,
   bf(y | cens(ycens, y2 = y2) ~ 1),
   data,
   prior(normal(0, 1), class = Intercept),
   car1 = FALSE,
   save_warmup = FALSE,
   chains = 3
 )
 loo_cv(data, fit, draw_ids = 1:3000, car1 = FALSE)

bentrueman/bgamcar1 documentation built on July 6, 2024, 11:16 p.m.