Limiting ESD for a weighted sample covariance matrix
In the authors' words
We study the empirical spectral distribution (ESD) of the random matrix , where is a rank orthogonal projector on , are i.i.d. standard Gaussian vectors, and , . We here assume that for some function discontinuous at . We show that the asymptotic ESD is characterized by a Marchenko--Pastur equation, independently of . We also show that the result extends to sub Gaussian random vectors and to the case where the projector is replaced with a more general weight . Last, in the Gaussian case, we study the emergence of outliers due to the weighting.
Main resultThe abstract does not state a limitation.
Appeared: Friday, October 30. arXiv. Preprint, not yet peer-reviewed.