Can someone please explain how regularization helps to shrink the " less important " features to zero ? As far as I know , Regularization only penalizes the weights of ALL the features to get them closer to 0 , so that they don't overfit to the training data . But how does it help in feature selection or deal with multicollinearity ?

Can someone please explain this in detail in a more intuitive manner ?

  • $\begingroup$ Does this answer your question? stats.stackexchange.com/q/74542/276972 $\endgroup$ Jul 23, 2020 at 7:14
  • $\begingroup$ If you haven't take your time to go through these videos, helped me a lot and they are described in detail and very intuitive. $\endgroup$
    – Thomas
    Jul 23, 2020 at 8:04