Accession Number : AD0646839

Title :   THE GENERALIZED VARIANCE: TESTING AND RANKING PROBLEMS,

Corporate Author : CHICAGO UNIV ILL

Personal Author(s) : Eaton,Morris L.

Report Date : 1967

Pagination or Media Count : 8

Abstract : The note shows that, for a sample from a multivariate normal distribution, the density function of the sample generalized variance possesses a monotone likelihood ratio. This result is used to construct a uniformly most powerful invariant test for a testing problem concerning the population generalized variance. Also, the result is applied to the problem of ranking multivariate normal populations according to the size of their generalized variances. (Author)

Descriptors :   (*STATISTICAL TESTS, *MULTIVARIATE ANALYSIS), PROBABILITY DENSITY FUNCTIONS, RANDOM VARIABLES, MONTE CARLO METHOD, DECISION THEORY, MINIMAX TECHNIQUE

Subject Categories : Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE