restriktor (version 0.6-10)

calculate_IC_weights: Calculating IC weights based on IC values (AIC, ORIC, GORIC(A), BIC, SIC, ...)

Description

This function transforms IC values into IC weights: IC values denote the ordering of hypotheses/models, while IC weights quantify the relative strength of hypotheses/models.

Usage

calculate_IC_weights(IC, hypo_names = NULL, use_scientific = TRUE)
calc_ICweights(IC, hypo_names = NULL, use_scientific = TRUE)
# S3 method for goric_ICw
print(x, digits = max(3, getOption("digits") - 4), ...)

Value

IC weights, which quantify the relative strength of hypotheses/models.

Arguments

IC

A vector or one-column matrix with information criteria (AIC, ORIC, GORIC(A), BIC, SIC, ...) values of length 'NrHypos', where 'NrHypos' stands for the number of hypotheses/ models.

x

an object of class con_goric.

hypo_names

Optional. Vector containing 'NrHypos' characters which will be used for labeling the hypothesis. Default: H1, H2, ...

use_scientific

If TRUE (default), the IC weights and ratio of IC weights will be formatted using scientific notation. If FALSE, standard numeric formatting is used.

digits

the number of significant digits to use when printing.

...

no additional arguments for now.

Examples

IC <- c(1,2,3)
calculate_IC_weights(IC)


## PT weights 
# This examples shows how to calculate PT weights.
# Notably, one is interested in PT weights when the log-likelihood for two or more
# hypotheses are (approximately) equal.
# Then, the comparison between those hypotheses is solely based on the PT values.
# The IC weights will then equal the PT weights.
# In that case, there is support for the overlap (boundary) of these hypotheses.
# Thus, when the IC weights equal the PT weights for a (sub)set of hypotheses,
# then there is support for the overlap (boundary) of these hypotheses.

y <- rnorm(30)
group <- factor(rep(c("A","B","C"), each = 10))
fit.lm <- lm(y ~ -1 + group)
est <- coef(fit.lm)
VCOV_est <- vcov(fit.lm)
H1 <- "groupA < groupB < groupC"
results <- goric(est, VCOV = VCOV_est, hypotheses = list(H1), 
                 comparison = "complement", type = "gorica")
calculate_IC_weights(results$result[,3])