In this paper, we introduce a proximal-proximalmajorization-minimization (PPMM) algorithm for nonconvex rank regression problems. The basic idea of the algorithm is to apply the proximalmajorization-minimization alg...
详细信息
In this paper, we introduce a proximal-proximalmajorization-minimization (PPMM) algorithm for nonconvex rank regression problems. The basic idea of the algorithm is to apply the proximalmajorization-minimizationalgorithm to solve the nonconvex problem with the inner subproblems solved by a sparse semismooth Newton (SSN) method based proximal point algorithm (PPA). It deserves mentioning that we adopt the sequential regularization technique and design an implementable stopping criterion to overcome the singular difficulty of the inner subproblem. Especially for the stopping criterion, it plays a very important role for the success of the algorithm. Furthermore, we also prove that the PPMM algorithm converges to a stationary point. Due to the Kurdyka-Lojasiewicz (KL) property of the problem, we present the convergence rate of the PPMM algorithm. Numerical experiments demonstrate that our proposed algorithm outperforms the existing state-of-the-art algorithms.
暂无评论