English

On the Noise-Information Separation of a Private Principal Component Analysis Scheme

Information Theory 2018-01-12 v1 math.IT

Abstract

In a survey disclosure model, we consider an additive noise privacy mechanism and study the trade-off between privacy guarantees and statistical utility. Privacy is approached from two different but complementary viewpoints: information and estimation theoretic. Motivated by the performance of principal component analysis, statistical utility is measured via the spectral gap of a certain covariance matrix. This formulation and its motivation rely on classical results from random matrix theory. We prove some properties of this statistical utility function and discuss a simple numerical method to evaluate it.

Keywords

Cite

@article{arxiv.1801.03553,
  title  = {On the Noise-Information Separation of a Private Principal Component Analysis Scheme},
  author = {Mario Diaz and Shahab Asoodeh and Fady Alajaji and Tamás Linder and Serban Belinschi and James Mingo},
  journal= {arXiv preprint arXiv:1801.03553},
  year   = {2018}
}

Comments

Submitted to the International Symposium on Information Theory (ISIT) 2018