Methods to calculate information criteria for
SVC_mle objects. Currently, two are supported: the conditional
Akaike's Information Criteria \(cAIC = -2*log-likelihood + 2*(edof + df)\)
and the Bayesian Information Criteria \(BIC = -2*log-likelihood + log(n) * npar\).
Note that the Akaike's Information Criteria is of the corrected form, that
is: \(edof\) is the effective degrees of freedom which is derived as the
trace of the hat matrices and df is the degree of freedoms with respect to
mean parameters.
# S3 method for SVC_mle
BIC(object, ...)# S3 method for SVC_mle
AIC(object, conditional = "BW", ...)
SVC_mle object
further arguments
string. If conditional = "BW", the
conditional AIC is calculated.
numeric, value of information criteria