Multimodel Inference Understanding AIC and BIC in Model Selection  

Oleh:

Burnham, Kenneth P ; Anderson, David R.

Jenis: Article from Bulletin/Magazine
Dalam koleksi: Sociological Methods and Research vol. 33 no. 02 (Nov. 2004), page 261.
Topik: AIC; BIC; model averaging; model selection; multimodel inference

Isi artikel

The model selection litelatur had been generally poor at reflecting the deep foundations of the akaike information criteria (AIC) and the making appropriate comparison to the Bayesian information criterion (BIC). there is clear philosophy, a sound criterion based in information theory, and rigorous statistical foundation for AIC. AIC can be justified as Bayesian using a "savvy" prior on the models that is the function of sample size of the number of models parameters. furthermore, BIC can be derived as a non Bayisian result. Therefor, arguments about using AIC versus BIC for model selection cannot be from a Bayes versus frequentist perspective.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali