Gamma Distribution and Partial Sufficiency of Polynomials,
ARMY FOREIGN SCIENCE AND TECHNOLOGY CENTER CHARLOTTESVILLE VA
Pagination or Media Count:
Considerable attention has been devoted currently in the statistical literature to problems connected with the characterization of probability distribution by various statistical properties. These can be used to construct interesting nonparametric criteria, and also are useful in the theory of linear models applied in biometrics. In addition to this the characterization problems lead to interesting classes of differential and integral equations. The paper considers the problem of characterizing the gamma distribution with an arbitrary integral or nonintegral number of degrees of freedom by the constant property of the regression of similar statistics on the sufficient statistic and partial sufficiency.
- Statistics and Probability