Online Hyperparameter-Free Sparse Estimation Method

Abstract

In this paper we derive an online estimator for sparse parameter vectors which, unlike the LASSO approach, does not require the tuning of any hyperparameters. The algorithm is based on a covariance matching approach and is equivalent to a weighted version of the square-root LASSO. The computational complexity of the estimator is of the same order as that of the online versions of regularized least-squares (RLS) and LASSO. We provide a numerical comparison with feasible and infeasible implementations of the LASSO and RLS to illustrate the advantage of the proposed online hyperparameter-free estimator.

References in corpus (9)

• Pathwise coordinate optimization
• Square-Root Lasso: Pivotal Recovery of Sparse Signals via Conic Programming
• Optimally Tuned Iterative Reconstruction Algorithms for Compressed Sensing
• Online Sparse System Identification and Signal Reconstruction using Projections onto Weighted ${\ell }_1$ Balls
• Weighted SPICE: A Unifying Approach for Hyperparameter-Free Sparse Estimation
• Dynamic Updating for L1 Minimization
• A variational Bayes framework for sparse adaptive estimation
• A Note on the SPICE Method
• On Phase Transition of Compressed Sensing in the Complex Domain

Cited by in corpus (6)

• Tuned Regularized Estimators for Linear Regression via Covariance Fitting
• Robust Prediction when Features are Missing
• A Block Alternating Optimization Method for Direction-of-Arrival Estimation with Nested Array
• Generalized Sparse Covariance-based Estimation
• Learning sparse linear dynamic networks in a hyper-parameter free setting
• Learning Sparse Graphs for Prediction and Filtering of Multivariate Data Processes