On the Asymptotic Properties of SLOPE
On the Asymptotic Properties of SLOPE
Abstract Sorted L-One Penalized Estimator (SLOPE) is a relatively new convex optimization procedure for selecting predictors in high dimensional regression analyses. SLOPE extends LASSO by replacing the L 1 penalty norm with a Sorted L 1 norm, based on the non-increasing sequence of tuning parameters. This allows SLOPE to adapt …