Hey,
in the lecture notes 3d) we are asked
I can picture some intuitive examples but can't get a complete "true" answer....
Thanks in advance ! :)
I suppose that, since L1 regularisation comes up with a sparse solution (some of the weights will be 0), it would be useful when you have to do feature selection. I'd appreciate if someone can confirm this :)
This question has been answered before: http://oknoname.herokuapp.com/forum/topic/229/l1-regularization-for-features-vs-l2/#c2
L1-L2 REG - Which is better ?
Hey,
in the lecture notes 3d) we are asked
I can picture some intuitive examples but can't get a complete "true" answer....
Thanks in advance ! :)
I suppose that, since L1 regularisation comes up with a sparse solution (some of the weights will be 0), it would be useful when you have to do feature selection. I'd appreciate if someone can confirm this :)
3
This question has been answered before:
http://oknoname.herokuapp.com/forum/topic/229/l1-regularization-for-features-vs-l2/#c2
2
Add comment