The Tracy-Widom law for the largest eigenvalue of F type matrices (Q309724)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The Tracy-Widom law for the largest eigenvalue of F type matrices
scientific article

    Statements

    The Tracy-Widom law for the largest eigenvalue of F type matrices (English)
    0 references
    0 references
    0 references
    0 references
    7 September 2016
    0 references
    This paper studies asymptotic laws of the largest eigenvalues of the equation \(\det (\lambda A_p+B_p)=0\) with \(A_p=\frac{1}{n}YY^*\) and \(B_p=\frac{1}{n}XX^*\) when the sizes of the matrices goes to infinity. A Tracy-Widom law for the largest eigenvalue is derived under moments conditions for the various matrices, but no invertibility condition is assumed. Though the results are not really a surprise for people in random matrix theory, the main interest of the present article lies in the large span of applications in statistics that are properly detailed by the authors. The redaction being oriented towards applications, the article remains concise and provides numerical simulations illustrating the main results. On the other hand, important parts of the proof of the main theorem are missing and are only provided in a supplementary article. Thus, the article is well suited for people interested in applications of the theorem or experts in the field but would be ill-advised for mathematicians interested in the elements of the proof.
    0 references
    random matrices
    0 references
    Tracy-Widom distribution
    0 references
    sample covariance matrix
    0 references
    Fisher matrix
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references