"Linear convergence of a weighted randomized sparse Kaczmarz method"NGOUPEYOU TONDJI, LIONELThe randomized version of the Kaczmarz method is an iterative algorithm for solving systems of linear equalities and inequalities, that iteratively projects onto these constraints. The sparse version of it was recently proposed to recover sparse solutions. Assuming that the matrix is standardized, in this work we introduce a weighted sampling variant of the randomized sparse Kaczmarz method based on the residuals, and prove its linear convergence in expectation with respect to the $l_2$-norm. |
| http://univie.ac.at/projektservice-mathematik/e/talks/NGOUPEYOU TONDJI_2021-06_Linear_convergence_of_a_weighted_randomized_sparse_Kaczmarz_method.pdf |
« back