正态随机矩阵的相关矩阵估计

The Correlation Matrix Estimator of the Normally Distributed Random Matrix

考虑某一由不相关的“信号”部分和“噪音”部分组成对象的多变量观察Xi,推广到一个 p×n随机矩阵X ,它服从均值为相关矩阵M、协方差矩阵为Γ (Σs+Σε) .给出了将“信号”与“噪音”

This paper considers the multivariable observation X i of an object composed of uncorrelated parts of “signals”and “noise”,and extends it to a p×n random matrix which is normally distributed with mean M and convariance Γ(Σ s +Σ ε ).The maximal likelihood estimator of Γ n after the separation of “signals”and “noise”is given.