Can you use AIC to compare non nested models?
The AIC does not require nested models. One of the neat things about the AIC is that you can compare very different models. However, make sure the likelihoods are computed on the same data. Beware of AIC values computed using conditional likelihoods because the conditioning may be different for different models.
Can BIC be used to compare non nested models?
If you want to compare two models that are not nested but are based on the same manifest variables, you can use BIC or AIC to compare the two models (samller values indicate better model fit; however, there is a descriptive comparison – you will not get a p-value for a difference test) – the critical point is that both …
Can a AIC be applied to a non nested model?
The AIC can be applied with non nested models. In fact, this is one of the most extended myths (misunderstandings?) about AIC. See: One thing you have to be careful about is to include all the normalising constants, since these are different for the different (non-nested) models:
When to use an extended model in AIC?
Thus, for any finite number of parametric models, we may always consider an extended model that will play the role of [ f (x|kθ) ] This suggests that AIC can be useful, at least in principle, for the comparison of models which are nonnested, i.e., the situation where the conventional log likelihood-ratio test is not applicable.” Akaike, Hirotugu.
Are there any other versions of the AIC?
Other versions of the AIC are discussed and compared in the following paper: For reference, a counterargument: Brian Ripley states in “Selecting amongst large classes of models” pp. 6-7 Crucial assumptions The models are nested (footnote: see the bottom of page 615 in the reprint of Akaike (1973)).
Which is an important observation about AIC?
“One important observation about AIC is that it is defined without specific reference to the true model [ f (x|kθ) ].