However, IRLS is computationally intensive and may not be suitable for massive data applications. In this letter, we propose a unified algorithmic framework for solving sparse regularization linear ...