ADMM Algorithmic Regularization Paths for Sparse and Large Scale Positive-Definite Covariance Matrix Estimation

Estimating sparse positive-definite covariance matrices in high dimensions has received extensive attention the past two decades. However, many existing algorithms are proposed for a single regularization parameter and little been paid to estimating over full range of parameters. In this paper we suggest compute paths through one-step approximation warm-starting Alternating Direction Method Multipliers (ADMM) algorithm, which quickly outlines sequence solutions at fine resolution. We demonstrate effectiveness computational savings our algorithm elaborative analysis simulated examples.