“…
regularization finds broad utilization in variable selection and feature extraction, which generates the sparsest solutions, yet these sparse solutions are often challenging to compute. [
30,31 ] To overcome this challenge,
regularization is introduced, but it does not exhibit sparsity as strong as L 0 regularization. [
32 ] It is indicated that
regularization assures the generation of sparser solutions in comparison to L 1 regularization, where the index 1/2 assumes a representative role.…”