Sensibilidad respecto a la función de pérdida en la teoría de la decisión.
Three methods are proposed for constructing reference prior densities for certain biparametric distribution families. These densities represent approximations to the Bayesian concept of noninformative distribution.
In this paper it is taken up a revision and characterization of the class of absolutely continuous elliptical distributions upon a parameterization based on the density function. Their properties (probabilistic characteristics, affine transformations, marginal and conditional distributions and regression) are shown in a practical and easy to interpret way. Two examples are fully undertaken: the multivariate double exponential distribution and the multivariate uniform distribution.
Page 1