The R Project SVN R

% File src/library/stats4/man/BIC.Rd
% Part of the R package, http://www.R-project.org
% Copyright 1995-2007 R Core Development Team
% Distributed under GPL 2 or later

\name{BIC}
%\docType{genericFunction}
\alias{BIC}
\alias{BIC,ANY-method}
\alias{BIC,logLik-method}
\title{Bayesian Information Criterion}
\description{
  This generic function calculates the Bayesian information criterion,
  also known as Schwarz's Bayesian criterion (SBC), for one or several
  fitted model objects for which a log-likelihood value can be obtained,
  according to the formula \eqn{-2 \mbox{log-likelihood} + n_{par}
    \log(n_{obs})}{-2*log-likelihood + npar*log(nobs)}, where
  \eqn{n_{par}}{npar}  represents the
  number of parameters and \eqn{n_{obs}}{nobs} the number of
  observations in the fitted model.
}
\usage{
BIC(object, \dots)
}
\arguments{
  \item{object}{An object of a suitable class for the BIC to be
    calculated - usually a \code{"logLik"} object or an object for which
    a \code{\link[=logLik-methods]{logLik}} method exists.
  }
  \item{\dots}{optionally more fitted model objects.}
}
\value{
  If just one object is provided, returns a numeric value
  with the corresponding BIC; if multiple objects are provided, returns
  a \code{\link{data.frame}} with rows corresponding to the objects and
  columns representing the number of parameters in the model (\code{df})
  and the BIC.
}
\references{
  Schwarz, G. (1978) "Estimating the Dimension of a Model",
  \emph{Annals of Statistics}, \bold{6}, 461--464.
}
\seealso{\code{\link{logLik-methods}}, \code{\link{AIC-methods}}}
\examples{
lm1 <- lm(Fertility ~ . , data = swiss)
AIC(lm1)
BIC(lm1)

## with two models:
lm1. <- update(lm1, . ~ . -Examination)
AIC(lm1, lm1.)
BIC(lm1, lm1.)

\testonly{
## 2 equivalent ways of calculating the BIC:
stopifnot(all.equal(AIC(lm1, k=log(nrow(swiss))), BIC(lm1)))
}
}
\keyword{models}