Information criteria application for model structure selection

Hi All,

I am working on Information criteria (IC) for model structure selection using identification tool box. There all types of IC are available AIC, FPE,BIC, nAIC,AICc.

• Can some one guide me which one to use when?
• How they are calculated for MIMO?

Looking forward for the answer...

Best regards

Answer 1

[Dinesh Yadav]

AIC penalizes complex models less, meaning that it may put more emphasis on model performance, on the training dataset and in turn, select more complex models.

BIC penalizes the model more for its complexity, meaning that more complex models will have a worse (larger) score and will, in turn, be less likely to be selected. It favours simple models.

The use in MIMO will be dependent on the research paper application you are referring to.

[Shafaq Gul]

Okay thankyou Dinesh...but Is it right that there is no special difference among them. But if the user wants more complexity they should use AIC otherwise should go for BIC.