L1 regularization (Lasso regularization): By adding the L1 norm of the weight vector to the loss function, the model can be made sparser, reducing the influence of unimportant features.
L2 regularization (Ridge regularization): Adding the L2 norm of the weight vector to the loss function can help prevent overfitting and make the weight vector’s values smoother.
Elastic Net regularization: by combining L1 and L2 regularization, it can achieve a better balance between sparsity and smoothness.
Group Lasso regularization: by grouping features and applying L1 regularization to each feature group, sparsity within the group is maintained.
Total Variation regularization is utilized in image processing to maintain the smoothness and edge information of an image.
TV-L1 regularization is a method that combines the characteristics of Total Variation regularization and L1 regularization, typically used in image restoration and denoising.
