CLME (version 2.0-12)

BIC.clme: Bayesian information criterion

Description

Calculates the Bayesian information criterion for objects of class clme.

Calculates the Akaike and Bayesian information criterion for objects of class clme.

Usage

# S3 method for clme
BIC(object, ..., k = log(nobs(object)/(2 * pi)))
# S3 method for summary.clme
BIC(object, ..., k = log(nobs(object)/(2 * pi)))

Arguments

object

object of class clme.

...

space for additional arguments.

k

value multiplied by number of coefficients

Value

Returns the Bayesian information criterion (numeric).

Details

The log-likelihood is assumed to be the Normal distribution. The model uses residual bootstrap methodology, and Normality is neither required nor assumed. Therefore the log-likelihood and these information criterion may not be useful measures for comparing models. For k=2, the function computes the AIC. To obtain BIC, set \(k = log( n/(2*pi) )\); which the method BIC.clme does.

See Also

CLME-package clme

CLME-package clme

Examples

# NOT RUN {
data( rat.blood )

cons <- list(order = "simple", decreasing = FALSE, node = 1 )
clme.out <- clme(mcv ~ time + temp + sex + (1|id), data = rat.blood , 
                 constraints = cons, seed = 42, nsim = 0)

BIC( clme.out )
BIC( clme.out, k=log( nobs(clme.out)/(2*pi) ) )


# }