A variational approach to stable principal component pursuit
Optimization and Control
2014-06-06 v1 Machine Learning
Abstract
We introduce a new convex formulation for stable principal component pursuit (SPCP) to decompose noisy signals into low-rank and sparse representations. For numerical solutions of our SPCP formulation, we first develop a convex variational framework and then accelerate it with quasi-Newton methods. We show, via synthetic and real data experiments, that our approach offers advantages over the classical SPCP formulations in scalability and practical parameter selection.
Cite
@article{arxiv.1406.1089,
title = {A variational approach to stable principal component pursuit},
author = {Aleksandr Aravkin and Stephen Becker and Volkan Cevher and Peder Olsen},
journal= {arXiv preprint arXiv:1406.1089},
year = {2014}
}
Comments
10 pages, 5 figures