Relaxed sparse eigenvalue conditions for sparse estimation via non-convex regularized regression

Non-convex regularizers usually improve the performance of sparse estimation in practice. To prove this fact, we study the conditions of sparse estimations for the sharp concave regularizers which ...