Computes Akaike difference and Akaike weights from an object of formal class “aic”.
# S4 method for aic
summary(object, which = c("AIC", "AICc"))
An object of formal class “aic”.
A character string indicating which information criterion is selected to compute Akaike difference and Akaike weights: either “AIC” or “AICc”.
The models are ordered according to AIC or AICc and 3 statistics are computed:
- the Akaike difference \(\Delta\): the change in AIC (or AICc) between successive (ordered) models,
- the Akaike weight \(W\): when \(r\) models are compared, \(W = e^{-0.5 * \Delta} / \sum_r{e^{-\frac{1}{2} * \Delta}}\),
- the cumulative Akaike weight \(cum.W\): the Akaike weights sum to 1 for the \(r\) models which are compared.
Burnham, K.P., Anderson, D.R., 2002. Model selection and multimodel inference: a practical
information-theoretic approach. New-York, Springer-Verlag, 496 p.
Hurvich, C.M., Tsai, C.-L., 1995. Model selection for extended quasi-likelihood models in small samples.
Biometrics, 51 (3): 1077-1084.
Examples in betabin and AIC in package stats.