摘要
C.R.Rao在[1]中都是在协方差矩阵为非奇异的情况下讨论的,作者把它们都推广到奇异的情形.作者在讨论用一个随机向且通过一个已知的随机向量中也得出一些结果,并肯定了C.R.Rao的一个猜测.
In[1],C.R.Rao discussed application of separation.theorems under nonsingularcovariance matrices.In this paper,we discusses them under singular covariance matrices.Some results are got when we discuss approximating a given random variable by anotherrandom variable.At last,a guess given by C.R.Rao is proved in this paper.
关键词
随机向量
逼近
奇异值分隔
定理
canonical correlation
canonical variate
unitanly invariant norm
singular value