I am reading up on model selection and ran into some intresting questions that I would like to understand to build intuition on the topic. My questions were:

  1. What are the appropriate ways of performing model selection?

  2. Are there any other alternatives for estimation? If so could you list these so I can read up on them?

  3. Can we say that the loss achieved on the training data is inappropriate and if so why is this the case?

  • $\begingroup$ You could read Regression Modeling Strategies by Frank Harrell, which covers all your questions (and many others). $\endgroup$ – Peter Flom Jul 3 '20 at 10:43