Constrained L1-minimization for Inverse (covariance) Matrix Estimation
A robust constrained L1 minimization method for estimating a large sparse inverse covariance matrix (aka precision matrix), and recovering its support for building graphical models. The computation uses linear programming.