0

I was reading this question here to understand the geomteric interpretation of OLS solution. I have seen this before, and I understand it: given that I take for granted that the minima can be interpreted as contour plots.

My question is how exactly I can relate the contour plots to this question. Why do we draw different contours for the objective and the penalty? If I can understand that, I can easily make a leap to why the intersection reveals the minimum (but any help with that would be greatly appreaciated as well).

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
learning
  • 465
  • 3
  • 16
  • Are you asking about regular OLS or penalized regression models such as lasso and ridge regression? – Matthew Gunn Jan 20 '17 at 04:33
  • Either or both. But I think I understood it now. The penalty term arises from a constraint term, so the contour is a constraint set. I just need to find a resource which shows how the constrained problem turns into this (unless you can point me to that?). – learning Jan 20 '17 at 04:36

0 Answers0