摘要
利用任意非负Borel函数的数学期望与随机向量的密度函数的关系,通过随机矩阵的变换,证明了左球分布定义的矩阵F和矩阵T仍然服从矩阵F分布和矩阵T分布,从而将矩阵F分布和矩阵T分布推广到左球分布类。这一结果扩大了椭球等高分布的应用范围。
Let A be a random matrix defined by left spherical family. In this paper, by using the relation between the expectation of non - negative random Borel function and density function, we prove that the F and T distributions of the matrix A are invariant under linear transformations. Consequently, we extend the F and T distributions of a random matrix to the left spherical family, which provides other applications of elliptically contoured distribution of random matrices.
出处
《南京邮电大学学报（自然科学版）》
2008年第5期64-67,共4页
Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition
基金
南京邮电大学校科研基金(NY207092)资助项目
关键词
椭球等高分布
矩阵T分布
矩阵F分布
变换
the elliptically contoured distribution
matrix T distribution
matrix F distribution
transformation
