1

I wanted to know how big are the regularization parameter values for ridge or lasso. I have seen most of the places generally using values like 0.1 or 0.01 but in some of my experiments the cross validation gives regularization parameters with values something like 8.6 and 172. Are these values fine or am I doing something wrong?

Shishir Pandey
  • 1,051
  • 2
  • 9
  • 11
  • Is this a real question? If yes, I recommend any standard text (ESL by Friedman-Hastie-Tibs is freely available online). If not, best question EVER! – gappy Sep 12 '13 at 02:47

1 Answers1

1

They're fine. Make sure to always normalize your data.

For a typical range, see also this question Which search range for determining SVM optimal C and gamma parameters?

Community
  • 1
ciri
  • 1,123
  • 9
  • 21