Tracy-Widom law for the extreme eigenvalues of sample correlation matrices

Abstract: Let the sample correlation matrix be

W = Y Y^{T}

, where

Y = (y_{i j})_{p, n}

with

y_{i j} = x_{i j} / s q r t s u m_{j = 1}^{n} x_{i j}^{2}

. We assume

x_{i j} : 1 l e q i l e q p, 1 l e q j l e q n

to be a collection of independent symmetric distributed random variables with sub-exponential tails. Moreover, for any

i

, we assume

x_{i j}, 1 l e q j l e q n

to be identically distributed. We assume

0 < p < n

and

p / n i g h t a r r o w y

with some

y i n (0, 1)

as

p, n i g h t a r r o w i n f t y

. In this paper, we provide the Tracy-Widom law (

T W_{1}

) for both the largest and smallest eigenvalues of

W

. If

x_{i j}

are i.i.d. standard normal, we can derive the

T W_{1}

for both the largest and smallest eigenvalues of the matrix

m a t h c a l R = R R^{T}

, where

R = (r_{i j})_{p, n}

with

,

.

Recommendations

Cited in

(36)

This page was built for publication: Tracy-Widom law for the extreme eigenvalues of sample correlation matrices