The graphical lasso: new insights and alternatives The graphical lasso  is an algorithm for learning the structure in an undirected Gaussian graphical model, using ℓ 1 regularization to control the number of zeros in the precision matrix Θ=Σ -1 [2, 11]. The R package glasso  is popular, fast, and allows one to efficiently build a path of models for different values of the tuning parameter. Convergence of glasso can be tricky; the converged precision matrix might not be the inverse of the estimated covariance, and occasionally it fails to converge with warm starts. In this paper we explain this behavior, and propose new algorithms that appear to outperform glasso. By studying the “normal equations” we see that, glasso is solving the dual of the graphical lasso penalized likelihood, by block coordinate ascent; a result which can also be found in . In this dual, the target of estimation is Σ, the covariance matrix, rather than the precision matrix Θ. We propose similar primal algorithms p-glasso and dp-glasso, that also operate by block-coordinate descent, where Θ is the optimization target. We study all of these algorithms, and in particular different approaches to solving their coordinate sub-problems. We conclude that dp-glasso is superior from several points of view.
Keywords for this software
References in zbMATH (referenced in 8 articles , 1 standard article )
Showing results 1 to 8 of 8.
- Chen, Lisha; Huang, Jianhua Z.: Sparse reduced-rank regression with covariance estimation (2016)
- Zhu, Hongxiao; Strawn, Nate; Dunson, David B.: Bayesian graphical models for multivariate functional data (2016)
- Forbes, Peter G.M.; Lauritzen, Steffen: Linear estimating equations for exponential families with application to Gaussian linear concentration models (2015)
- Fitch, A.Marie; Jones, M.Beatrix; Massam, Hélène: The performance of covariance selection methods that consider decomposable models only (2014)
- Fitch, A.Marie; Jones, Beatrix: The cost of using decomposable Gaussian graphical models for computational convenience (2012)
- Mazumder, Rahul; Hastie, Trevor: The graphical lasso: new insights and alternatives (2012)
- Zhao, Tuo; Liu, Han; Roeder, Kathryn; Lafferty, John; Wasserman, Larry: The huge package for high-dimensional undirected graph estimation in R (2012)
- Lu, Zhaosong: Adaptive first-order methods for general sparse inverse covariance selection (2010)