In regression models to predict chemical compounds' activities, fingerprints are often used as features. Should one normalize a fingerprint feature to be in the 0-1 range?

  • 2
    $\begingroup$ I've posted this question in a chat room with some machine learning experts, and there's a question about whether you're doing ordinary least-squares regression or maximum likelihood estimation procedures. Perhaps you can join that discussion and we can converge closer to an answer for you: chat.stackexchange.com/transcript/message/56950547#56950547 $\endgroup$ Feb 3 at 16:21
  • $\begingroup$ I think the question would benefit from some more detail on the sorts of compounds and features you are looking at. The question is definitely on-topic here, but if you don't wind up getting a response, you may also want to try AI SE or Cross Validated. $\endgroup$
    – Tyberius
    Jun 23 at 21:41