DCA for Sparse Quadratic Kernel-Free Least Squares Semi-Supervised Support Vector Machine

With the development of science and technology, more data have been produced. For many these datasets, only some labels. In order to make full use information in data, it is necessary classify them. this paper, we propose a strong sparse quadratic kernel-free least squares semi-supervised support vector machine (SSQLSS3VM), which add ℓ0norm regularization term sparse. An NP-hard problem arises since proposed model contains ℓ0 norm another nonconvex term. One important method for solving DC (difference convex function) programming. Therefore, first approximate by polyhedral function. Moreover, due existence nonsmooth terms, sGS-ADMM solve subproblem. Finally, empirical numerical experiments show efficiency algorithm.