Enhanced Lasso Recovery on Graph

TL;DR

This paper introduces a novel non-convex Lasso algorithm tailored for sparse signal recovery on graphs, outperforming standard methods through numerical experiments on benchmark datasets.

Contribution

It combines compressed sensing and graph Fourier analysis to develop a non-convex Lasso algorithm specifically for graph-structured data.

Findings

Outperforms standard convex Lasso in experiments

Effective on multiple benchmark graph datasets

Provides a new approach for sparse graph signal recovery

Abstract

This work aims at recovering signals that are sparse on graphs. Compressed sensing offers techniques for signal recovery from a few linear measurements and graph Fourier analysis provides a signal representation on graph. In this paper, we leverage these two frameworks to introduce a new Lasso recovery algorithm on graphs. More precisely, we present a non-convex, non-smooth algorithm that outperforms the standard convex Lasso technique. We carry out numerical experiments on three benchmark graph datasets.