best_models: Precision for the k Best-Performing Models

In danheck/MCMCprecision: Precision of Discrete Parameters in Transdimensional MCMC

Precision for the k Best-Performing Models

Description

Assesses the precision in estimating the ranking of the k best-performing models.

Usage

best_models(samples, k, ties.method = "min")

Arguments

samples	a matrix with posterior samples (one per row) for the model posterior probabilities (one model per column). Can be estimated using stationary with the argument summary = FALSE.
k	number of best-performing models to be considered
ties.method	a character string specifying how ties are treated, see rank

Examples

# a sequence of uncorrelated model indices
mult <- rmultinom(1000, 1, c(.05, .6, .15, .12, .08))
idx <- apply(mult, 2, which.max)
z <- letters[idx]
stat <- stationary(z, summary = FALSE)
best_models(stat, 3)

danheck/MCMCprecision documentation built on Nov. 13, 2022, 11:41 p.m.