Product of Two Random Matrices

Abstract

In this chapter, we shall consider the LSD of a product of two random matrices, one of them a sample covariance matrix and the other an arbitrary Hermitian matrix. This topic is related to two areas: The first is the study of the LSD of a multivariate F-matrix that is a product of a sample covariance matrix and the inverse of another sample covariance matrix, independent of each other. Multivariate F plays an important role in multivariate data analysis, such as two-sample tests, MANOVA (multivariate analysis of variance), and multivariate linear regression. The second is the investigation of the LSD of a sample covariance matrix when the population covariance matrix is arbitrary. The sample covariance matrix under a general setup is, as mentioned in Chapter 3, fundamental in multivariate analysis.