English

Sharp Sufficient Conditions on Exact Sparsity Pattern Recovery

Information Theory 2009-10-13 v3 math.IT

Abstract

Consider the nn-dimensional vector y=X\be+\ey=X\be+\e, where \beRp\be \in \R^p has only kk nonzero entries and \eRn\e \in \R^n is a Gaussian noise. This can be viewed as a linear system with sparsity constraints, corrupted by noise. We find a non-asymptotic upper bound on the probability that the optimal decoder for β\beta declares a wrong sparsity pattern, given any generic perturbation matrix XX. In the case when XX is randomly drawn from a Gaussian ensemble, we obtain asymptotically sharp sufficient conditions for exact recovery, which agree with the known necessary conditions previously established.

Keywords

Cite

@article{arxiv.0910.0456,
  title  = {Sharp Sufficient Conditions on Exact Sparsity Pattern Recovery},
  author = {Kamiar Rahnama Rad},
  journal= {arXiv preprint arXiv:0910.0456},
  year   = {2009}
}

Comments

submitted to IEEE Trans. on Information Theory

R2 v1 2026-06-21T13:53:33.422Z