choshin84 / learning_memo

personal learning memo

0 stars 0 forks source link

AIC/BIC vs. CV #38

Open choshin84 opened 5 years ago

choshin84 commented 5 years ago

Tweet summary

AIC/BIC is criteria to show how appropriate the model is, NOT accurate on the prediction/inference. (Maybe useful in case of small sample size situation...)

Useful link

https://news.ycombinator.com/item?id=16094822

choshin84 commented 5 years ago

AIC is to select model based on how good is the prediction will be. データ解析のための統計モデリング入門p.85