Extract Information Criteria from a List of Parametric Survival Models

Extracts either the Akaike Information Criterion (AIC) or the Bayesian Information Criterion (BIC) from each flexsurvreg object in a named list, returning a named numeric vector for use with compute_weights.

Usage

extract_ic(models, criterion = c("AIC", "BIC"))

Arguments

models

A named list of fitted flexsurvreg objects, typically the output of fit_models.

criterion

Character scalar: "AIC" (default) or "BIC".

Value

A named numeric vector of information criterion values, one per model.

Examples

# \donttest{
t <- 0:5
s <- c(1, 0.549, 0.440, 0.394, 0.363, 0.341)
ipd   <- prep_ipd(t, s)
mods  <- fit_models(ipd[-1, ])
extract_ic(mods, "AIC")
#>      exp  weibull gompertz   llogis    lnorm    gamma 
#> 3.375273 5.375252 5.356812 5.357385 5.331519 5.374326 
extract_ic(mods, "BIC")
#>       exp   weibull  gompertz    llogis     lnorm     gamma 
#> 0.7756162 0.1759382 0.1574987 0.1580710 0.1322055 0.1750123 
# }