How can shrinkage methods help you improve model fit?
Asked
Active
Viewed 18 times
0
-
Technically, shrinkage methods make you fit the data slightly worse, not better. You bias your estimates to be less extreme in hopes that they become less variable. Sometimes, this has the effect of increasing some performance metric. See the James-Stein estimator. – Demetri Pananos May 11 '21 at 20:42