Edinburgh Research Explorer

Spectral properties of empirical covariance matrices for data with power-law tails

Research output: Contribution to journalArticle

Related Edinburgh Organisations

Original languageEnglish
Article number041129
Pages (from-to)-
Number of pages7
JournalPhysical Review E - Statistical, Nonlinear and Soft Matter Physics
Volume74
Issue number4
DOIs
StatePublished - Oct 2006

Abstract

We present an analytic method for calculating spectral densities of empirical covariance matrices for correlated data. In this approach the data is represented as a rectangular random matrix whose columns correspond to sampled states of the system. The method is applicable to a class of random matrices with radial measures including those with heavy (power-law) tails in the probability distribution. As an example we apply it to a multivariate Student distribution.

    Research areas

  • FINANCIAL CORRELATION-MATRICES, DIMENSIONAL RANDOM MATRICES, PORTFOLIO OPTIMIZATION, LEVY MATRICES, NOISE, EIGENVALUES, SIGNAL

ID: 1217429