Caesar AI Atlas

L~1~ Regularization

Caesar AI Atlas Definition

L1 regularization adds a penalty proportional to the absolute values of model weights. It encourages sparsity by pushing some weights to zero, which can support feature selection and simpler, more interpretable models.

Other Definitions

L~1~ Regularization Source

A type of regularization that penalizes weights in proportion to the sum of the absolute value of the weights. L~1~ regularization helps drive the weights of irrelevant or barely relevant features to exactly 0 . A feature with a weight of 0 is effectively removed from the model. Contrast with L~2~ regularization.

Concept Comparisons

Related Terms