questions-notes-2016-07-05-ch6.txt

For ridge regression:
l2 norm - why would I tune that instead of lambda?

Scaling of predictors!

What is beta again?  It's the vector of coefficients for a regression, and beta-hat is the estimate.  (Right?)

Why use selection over lasso?  Because you can use some selected variables in a nonlinear model, but lasso embeds the linear regression.