Number of Features

[dgouveiajr]

Hi all,

First of all, congrats for the course. It is amazing the content that you guys are making. My question is regarding the number of samples that may underfit or that are optimal for using the linear regression. Is it there some standard value on this or depends of the dataset?

[glemaitre]

It is really dataset-dependent and even problem specific. Usually, more features induced a greater chance to overfit (and thus the need for regularization).