Skip to contents

Summarizes IC differences and weights to identify plausible models (models with highest empirical support).

Usage

summaryIC(x, delta = 2)

Arguments

x

A data frame with one row per model, and one column giving an information-theoretic index (AIC or BIC), such as returned by AIC or BIC with several models.

delta

The difference threshold to identify models with similar support.

Value

The input data frame with additional columns delta giving IC differences with the best model, best identifying the best model(s) (* for the absolute best model, + for models within the threshold), and omega giving AIC weights.

Author

Mathieu Basille <\email{mathieu@basille.org}>

Examples

## Prepare two models:
lm1 <- lm(Fertility ~ . , data = swiss)
lm2 <- update(lm1, . ~ . -Examination)

## Check AIC:
AIC(lm1, lm2)
#>     df      AIC
#> lm1  7 326.0716
#> lm2  6 325.2408

## Summary of AIC and BIC:
summaryIC(AIC(lm1, lm2))
#>     df      AIC     delta     omega best
#> lm1  7 326.0716 0.8307244 0.3976271    +
#> lm2  6 325.2408 0.0000000 0.6023729    *
summaryIC(BIC(lm1, lm2))
#>     df      BIC    delta     omega best
#> lm1  7 339.0226 2.680872 0.2074384     
#> lm2  6 336.3417 0.000000 0.7925616    *