Is there a main reference/publication suggesting to use cross-validation in order to choose the regularization parameter in Lasso?
1
-
1$\begingroup$ The original Lasso paper by Tibshirani discusses using cross-validation to choose the regularisation parameter in Section 4. $\endgroup$– David RCommented Sep 29, 2016 at 15:40
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
You can try sec. 6.2.3 of James/Witten/Hastie/Tibshirani, Introduction to Statistical Learning
... implementing ridge regression and the lasso requires a method for selecting a value for the tuning parameter λ ... Cross-validation provides a simple way to tackle this problem. We choose a grid of λ values, and compute the cross-validation error for each value of λ ... We then select the tuning parameter value for which the cross-validation error is smallest. (p. 227)
The canonical text for Lasso is probably Hastie/Tibshirani/Friedman, Elements of Statistical Learning, but ISLR is a good place to start for non-PhD level.
