compare_BICw: Compare BIC weights for a set of models.

In hstojic/cogmod: Useful functions for cognitive modeling

Description

With BIC weights instead of raw BIC values one can get the sense of magnitude how much more likely (in terms of Kullback-Leibler discrepancy) one model is than some other model that has also been fitted. To get such a likelihood we divide the BIC weight of one model with the BIC weight of the other fitted model. See Wagenmakers and Farrel (2004) for details.

Usage

compare_BICw(BICweights, model_names = NA)

Arguments

BICweights	A numeric vector of BIC weigths.
model_names	Optional character vector with names of the models.

Value

A data frame where each cell represents a likelihood of the model denoted by the row name with respect to the model denoted by the cell's column name. Order of the models from the input to the function is preserved.

Examples

# example from Wagenmakers and Farrel (2004) article
BIC <- c(204, 202, 206, 206, 214)
BICweights <- BICw(BIC)$BICweights
# comparing BIC weights
compare_BICw(BICweights)

hstojic/cogmod documentation built on May 17, 2019, 6:16 p.m.