Akaike Information Criterion

DESCRIPTION:

This generic function calculates the Akaike information criterion for one or several fitted model objects for which a log-likelihood value can be obtained, according to the formula -2*log-likelihood + 2*npar, where npar represents the number of parameters in the fitted model. When comparing fitted objects, the smaller the AIC, the better the fit.

USAGE:

AIC(object, ...) 

REQUIRED ARGUMENTS:

object
a fitted model object, for which there exists a logLik method to extract the corresponding log-likelihood, or an object inheriting from class logLik.

OPTIONAL ARGUMENTS:

...
optional fitted model objects.

VALUE:

if just one object is provided, returns a numeric value with the corresponding AIC; if more than one object are provided, returns a data.frame or bdFrame with rows corresponding to the objects and columns representing the number of parameters in the model ( df) and the AIC.

WARNING:

Although the Akaike Information Criterion is a measure of goodness of fit for a more general class of models, this function is part of the Non-Linear Mixed Effects library in S-PLUS (nlme3) and currently there are methods developed only for functions within that library. To learn what functions have AIC methods in your current implementation of S-PLUS, enter the command methods(AIC) at the S-PLUS prompt.

REFERENCES:

Sakamoto, Y., Ishiguro, M., and Kitagawa G. (1986) "Akaike Information Criterion Statistics", D. Reidel Publishing Company.

SEE ALSO:

, ,

EXAMPLES:

fm1 <- lm(distance ~ age, data = Orthodont) # no random effects 
fm2 <- lme(distance ~ age, data = Orthodont) # random is ~age 
AIC(fm1, fm2)