ICOMP: ICOMP

In wrathematics/icomp: ICOMP criterion

Description

Generic function for calculating Bozdogan's information complexity measure (ICOMP). ICOMP is a likelihood-based criterion similar to Akaike's Information Criterion (AIC). As with AIC, a smaller ICOMP value corresponds to a better model.

Usage

## S3 method for class 'lm'
ICOMP(object, ..., cplex = "C1")

ICOMP(object, ...)

Arguments

object	A model for which there is a method for stats::logLik(). Currently supports lm objects.
...	Additional arguments.
cplex	Complexity measure, as a string. Valid options are "C0", "C1", "C1F", and "COND", the latter being condition number.

Value

The ICOMP value for the given model.

Author(s)

Drew Schmidt and Jake Ferguson

References

Bozdogan, H. Haughton, D.M.A (1998). Information complexity criteria for regression models. Computation Statistics & Data Analysis 28: 51-76

Examples

## Not run: 
### Modified from the example in ?AIC
lm1 <- lm(Fertility ~ . , data = swiss)
lm2 <- update(lm1, . ~ . -Examination)

icomp::ICOMP(lm1, lm2)

## End(Not run)

wrathematics/icomp documentation built on May 4, 2019, 9:49 a.m.