What is common in these concepts and what do they differ?
Do I understand correctly that with L1 regularization, part of the signs are completely discarded, and with L2 regularization, signs are smoothed so that due to one sign did not “fit” the others? And L1 and L2 norm, are it just different options in order to calculate the distance?
Answer 1, Authority 100%
General formula for L p norm :
L 1 norm / Distance of city blocks :
l 2 norm / Euclidean metric :
For regularization to the error formula, the corresponding norm is added – with L 1 regularization Add L 1 norm and L 2 Norm in the case of L 2 regularization :
