Distribution of the largest root of a matrix for Roy's test in multivariate analysis of variance

Abstract

Let $\mathbf{X},\mathbf{Y}$ denote two independent real Gaussian $\mathsf{p}\times \mathsf{m}$ and $\mathsf{p}\times \mathsf{n}$ matrices with $\mathsf{m},\mathsf{n}\ge \mathsf{p}$, each constituted by zero mean i.i.d. columns with common covariance. The Roy's largest root criterion, used in multivariate analysis of variance (MANOVA), is based on the statistic of the largest eigenvalue, ${\Theta }_1$, of ${(\mathbf{A}+\mathbf{B})}^{-1}\mathbf{B}$, where ${\mathbf{A}=\mathbf{XX}}^T$ and ${\mathbf{B}=\mathbf{YY}}^T$ are independent central Wishart matrices. We derive a new expression and efficient recursive formulas for the exact distribution of ${\Theta }_1$. The expression can be easily calculated even for large parameters, eliminating the need of pre-calculated tables for the application of the Roy's test.

References in corpus (2)

• Multivariate analysis and Jacobi ensembles: largest eigenvalue, Tracy--Widom limits and rates of convergence
• Approximate null distribution of the largest root in multivariate analysis

Cited by in corpus (9)

• On the probability that all eigenvalues of Gaussian, Wishart, and double Wishart random matrices lie within an interval
• Permutation Inference for Canonical Correlation Analysis
• On the Distribution of an Arbitrary Subset of the Eigenvalues for some Finite Dimensional Random Matrices
• Roy's Largest Root Test Under Rank-One Alternatives
• Asymptotic regime for impropriety tests of complex random vectors
• A Tracy-Widom Empirical Estimator For Valid P-values With High-Dimensional Datasets
• Largest Finite Root of Identity-Scale Doubly Singular Beta Type II Ensembles
• Computation of the Expected Euler Characteristic for the Largest Eigenvalue of a Real Non-central Wishart Matrix
• Expressing the largest eigenvalue of a singular beta F-matrix with heterogeneous hypergeometric functions