No label defined (Q4938009)

The Marchenko-Pastur distribution appeared in [\textit{V. A. Marchenko} and \textit{L. A. Pastur}, Math. USSR, Sb. 1, 457-483 (1968); translation from Mat. Sb., n. Ser. 72(114), 507-536 (1967; Zbl 0152.16101)] as the limiting mean spectral density of certain random matrices. The Wishart matrix is a particular, but important example. \textit{F. Oravecz and D. Petz} [Acta Sci. Math. 63, No. 3/4, 383-395 (1997; Zbl 0889.15006)] gave an elementary combinatorial proof concerning the mean spectral density of the above mentioned matrix. The limit theorem the author treats in this paper concerns the symmetric matrix \((TT^{\text{t}})/n\), where t stands for the transpose and \(T=T(n)\) is a random matrix with identically distributed independent standardized elements, where all moments are bounded. It is shown that the empirical spectral density of this matrix converges in distribution almost surely to a variant of the Marchenko-Pastur distribution as \(n\to \infty\). The proof is based on the method of moments. A more general result can be found by \textit{K. W. Wachter} [Ann. Probab. 6, 1-18 (1978; Zbl 0374.60039)] concerning sample matrices of independent elements. However the proof of the special case given here is simpler (only based on elementary combinatorial methods).