Welcome to GGLasso’s documentation!

GGLasso (stands for General Graphical Lasso) is a Python package for estimating sparse (or sparse plus low rank) inverse covariance matrices. Moreover, it contains solvers and model selection procedures for Multiple Graphical Lasso problems such as Group and Fused Graphical Lasso.

The package is available on Github .

Contents:

• Getting Started
  • Installation
• Examples Gallery
• Mathematical Description
  • Single Graphical Lasso problems (SGL)
  • SGL with latent variables
  • Multiple Graphical Lasso problems (MGL)
  • MGL with latent variables
  • GGL - the nonconforming case
  • Functional Graphical Lasso
  • Optimization algorithms
  • References
• Using the Problem Object
  • Class glasso_problem
  • Other methods of glasso_problem
  • Class GGLassoEstimator
  • Model selection
• Algorithms Overview
  • SGL solver
  • MGL solver
  • Nonconforming GGL solver
  • Functional Graphical Lasso solver
  • Further remarks on proximal operators
• Detailled Solver Documentation
  • SGL
  • MGL
  • GGL nonconforming
  • Functional SGL