logLik
, according to the formula
log-likelihood +
npar*log(nobs), where npar represents the number of
parameters and nobs the number of observations in the
fitted model. When comparing fitted objects, the smaller the BIC, the
better the fit.
BIC(object)
logLik
, usually
resulting from applying a
logLik
method to a fitted model
object.
Schwarz, G. (1978) "Estimating the Dimension of a Model", Annals of Statistics, 6, 461-464.
fm1 <- lm(distance ~ age, data = Orthodont) BIC(logLik(fm1))