Preprint
Jul 2026
Estimation of multiple precision matrices under shared support with heterogeneous edge strengths
The Multiplicative Graphical Lasso (Mglasso) is introduced, a method for jointly estimating precision matrices across multiple Gaussian graphical models under a shared sparsity constraint that establishes the local strict convexity of the objective function and provides rigorous high-dimensional consistency guarantees.
S. Bhowal, Debashis Paul, Gopal K. Basak et al.
· 0 citations