Eigenvalues and eigenvectors of large-dimensional sample covariance matrices
Jack W. Silverstein · Contemporary mathematics - American Mathematical Society · 1986
Limit theorems will be reviewed on the eigenvalues of a of class sample covariance matrices where the number oi v.iior-iJrpf.r and the vector dimension are on the same order of magnitude. iite main resu'lt states that the empirical distribution function oi the.ig.nvulues converges almost surely to a nonrandom distribution tunci.iJn, ur the dimension.approaches ihtinity, The author wlii then pr.ienl'r'it results in.describing tl,. behavior of the eigenvectors of these matrices. The results suggest similarlty between the-measure on the apfropriate orthogonal group induced by the matrix of eigenvectors and Haar measure.