Does the ℓ_1-norm Learn a Sparse Graph under Laplacian Constrained Graphical Models?

by   Jiaxi Ying, et al.

We consider the problem of learning a sparse graph under Laplacian constrained Gaussian graphical models. This problem can be formulated as a penalized maximum likelihood estimation of the precision matrix under Laplacian structural constraints. Like in the classical graphical lasso problem, recent works made use of the ℓ_1-norm regularization with the goal of promoting sparsity in Laplacian structural precision matrix estimation. However, we find that the widely used ℓ_1-norm is not effective in imposing a sparse solution in this problem. Through empirical evidence, we observe that the number of nonzero graph weights grows with the increase of the regularization parameter. From a theoretical perspective, we prove that a large regularization parameter will surprisingly lead to a fully connected graph. To address this issue, we propose a nonconvex estimation method by solving a sequence of weighted ℓ_1-norm penalized sub-problems and prove that the statistical error of the proposed estimator matches the minimax lower bound. To solve each sub-problem, we develop a projected gradient descent algorithm that enjoys a linear convergence rate. Numerical experiments involving synthetic and real-world data sets from the recent COVID-19 pandemic and financial stock markets demonstrate the effectiveness of the proposed method. An open source 𝖱 package containing the code for all the experiments is available at


page 6

page 15


Learning Graph Laplacian with MCP

Motivated by the observation that the ability of the ℓ_1 norm in promoti...

Efficient Graph Laplacian Estimation by a Proximal Newton Approach

The Laplacian-constrained Gaussian Markov Random Field (LGMRF) is a comm...

Structured Graph Learning Via Laplacian Spectral Constraints

Learning a graph with a specific structure is essential for interpretabi...

Sparse Graph Learning Under Laplacian-Related Constraints

We consider the problem of learning a sparse undirected graph underlying...

Algorithms for Learning Graphs in Financial Markets

In the past two decades, the field of applied finance has tremendously b...

Optimal designs for discrete choice models via graph Laplacians

In discrete choice experiments, the information matrix depends on the mo...

Total positivity in multivariate extremes

Positive dependence is present in many real world data sets and has appe...

Please sign up or login with your details

Forgot password? Click here to reset