Proximal gradient algorithm for convex optimization, using a diagonal +/- rank-1 norm. Uses special tricks to allow the use of a quasi-Newton methods.
Stephen Becker (2020). zeroSR1 (https://github.com/stephenbeckr/zeroSR1), GitHub. Retrieved .
Inspired by: NNLS and constrained regression, predictor-corrector algorithm, nnls, active set algorithm, newton's algorithm for nnls, MTRON, LARS algorithm, LBFGSB (L-BFGS-B) mex wrapper, mex interface for bound constrained optimization via ASA, nnls - Non negative least squares, Simple MATLAB example code and generic function to perform LASSO