aic {nadiv} | R Documentation |
Akaike Information Criterion
Description
Calculates AIC/AICc values, AIC differences, Likelihood of models, and model probabilities.
Usage
aic(logLik, fp, n = NULL)
Arguments
logLik |
A vector of model log-Likelihoods |
fp |
A vector containing the numbers of free parameters of each model included in the logLik vector |
n |
An optional vector of sample sizes for each model. Used to calculate AICc (small sample unbiased AIC). |
Details
Calculations and notation follows chapter 2 of Burnham and Anderson (2002).
Value
a list
:
- AIC
vector containing AIC/AICc (depending on value of
n
)- delta_AIC
vector containing AIC differences from the minimum AIC(c)
- AIClik
vector containing likelihoods for each model, given the data. Represents the relative strength of evidence for each model.
- w
Akaike weights.
Author(s)
References
Burnham, K.P. and D.R. Anderson. 2002. Model Selection and Multimodel Inference. A Practical Information-Theoretic Approach, 2nd edn. Springer, New York.
Examples
aic(c(-3139.076, -3136.784, -3140.879, -3152.432), c(8, 7, 8, 5))