On the largest eigenvalue of Wishart matrices with identity covariance when n, p and p/n tend to infinity

Abstract

Let X be a n*p matrix and l_1 the largest eigenvalue of the covariance matrix X^{*}*X. The "null case" where X_{i,j} are independent Normal(0,1) is of particular interest for principal component analysis. For this model, when n, p tend to infinity and n/p tends to gamma in (0,\infty), it was shown in Johnstone (2001) that l_1, properly centered and scaled, converges to the Tracy-Widom law. We show that with the same centering and scaling, the result is true even when p/n or n/p tends to infinity. The derivation uses ideas and techniques quite similar to the ones presented in Johnstone (2001). Following Soshnikov (2002), we also show that the same is true for the joint distribution of the k largest eigenvalues, where k is a fixed integer. Numerical experiments illustrate the fact that the Tracy-Widom approximation is reasonable even when one of the dimension is "small".

arXiv comment24 pages

References in corpus (2)

• Level-Spacing Distributions and the Airy Kernel
• Noise Dressing of Financial Correlation Matrices

Cited by in corpus (17)

• Beta ensembles, stochastic Airy spectrum, and a diffusion
• High Dimensional Statistical Inference and Random Matrices
• Distribution of the largest eigenvalue for real Wishart and Gaussian random matrices and a simple approximation for the Tracy-Widom distribution
• Limits of spiked random matrices I
• Convergence of the largest eigenvalue of normalized sample covariance matrices when p and n both tend to infinity with their ratio converging to zero
• Cross-Validation for Unsupervised Learning
• Application of Random Matrix Theory to Multivariate Statistics
• CLT for linear spectral statistics of normalized sample covariance matrices with the dimension much larger than the sample size
• Limiting Statistics of the Largest and Smallest Eigenvalues in the Correlated Wishart Model
• Universality in Complex Wishart ensembles: The 2 cut case
• Edgeworth Expansion of the Largest Eigenvalue Distribution Function of GUE Revisited
• Small deviations for beta ensembles
• Universality in Complex Wishart ensembles: The 1 cut case
• A Tracy-Widom Empirical Estimator For Valid P-values With High-Dimensional Datasets
• Testing for subsphericity when $n$ and $p$ are of different asymptotic order
• Approximation of Rectangular Beta-Laguerre Ensembles and Large Deviations
• On asymptotic expansion and CLT of linear eigenvalue statistics for sample covariance matrices when $N/M\to 0$