On the empirical spectral distribution for certain models related to sample covariance matrices with different correlations

Asymptotic distribution theory in statistics (62E20) Random matrices (algebraic aspects) (15B52) Random matrices (probabilistic aspects) (60B20) Strong limit theorems (60F15)

Abstract: Given

n, m i n m a t h b b N

, we study two classes of large random matrices of the form

mathcal{L}_n =sum_{alpha=1}^mxi_alpha mathbf{y}_alpha mathbf{y}_alpha ^Tquad ext{and}quad mathcal{A}_n =sum_{alpha =1}^mxi_alpha (mathbf{y}_alpha mathbf{x}_alpha ^T+mathbf{x}_alpha mathbf{y}_alpha ^T),

where for every

n

,

(x i_{a} l p h a)_{a} l p h a s u b s e t m a t h b b R

are iid random variables independent of

(m a t h b f x_{a} l p h a, m a t h b f y_{a} l p h a)_{a} l p h a

, and

(m a t h b f x_{a} l p h a)_{a} l p h a

,

(m a t h b f y_{a} l p h a)_{a} l p h a s u b s e t m a t h b b R^{n}

are two (not necessarily independent) sets of independent random vectors having different covariance matrices and generating well concentrated bilinear forms. We consider two main asymptotic regimes as

n, m (n) o i n f t y

: a standard one, where

m / n o c

, and a slightly modified one, where

m / n o i n f t y

and

m a t h b f E x i o 0

while

m m a t h b f E x i / n o c

for some

c g e 0

. Assuming that vectors

(m a t h b f x_{a} l p h a)_{a} l p h a

and

(m a t h b f y_{a} l p h a)_{a} l p h a

are normalized and isotropic "in average", we prove the convergence in probability of the empirical spectral distributions of

m a t h c a l L_{n}

and

m a t h c a l A_{n}

to a version of the Marchenko-Pastur law and so called effective medium spectral distribution, correspondingly. In particular, choosing normalized Rademacher random variables as

(x i_{a} l p h a)_{a} l p h a

, in the modified regime one can get a shifted semicircle and semicircle laws. We also apply our results to the certain classes of matrices having block structures, which were studied in [9, 21].

Recommendations

Cites work

Cited in

(4)

This page was built for publication: On the empirical spectral distribution for certain models related to sample covariance matrices with different correlations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5092965)