Maximum Likelihood Character of Distributions

Fan Jianqing,Fang Kaitai
DOI: https://doi.org/10.1007/bf02008374
1987-01-01
Acta Mathematicae Applicatae Sinica
Abstract:In this paper,some distributions in the family of those with invariance under orthogonaltransformations within an s-dimensional linear subspace are characterized by maximun likelihoodcriteria.Specially,the main result is:suppose P v is a projection matrix of a given s-dimensionalsubspace V,and x 1 ,…,x n are i.i.d.samples drawn from a population with a pdf f(x′P v x),wheref(·) is a positive and continuously differentiable function.Then P v (M n ) is the maximum likelihoodestimator of P v ifff(x)=c k exp(kx) (k>0),where M n =sum from i=1 to n x i x′ i ,P v (M n )=sum from i=1 to (?) (?) i (?)′ t ,λ 1 ,…,λ ? are the first s largest eigenvalues of matrix M n ,and(?) 1 ,…,(?) ? are their associated eigenvectors.
What problem does this paper attempt to address?