Inferences about the coefficient of correlation in the standard bivariate normal distribution

The study of the association between two random variables that have a joint normal distribution is of interest in applied statistics; for example, in statistical genetics. This article, targeted to applied statisticians, addresses inferences about the coefficient of correlation (ρ) in the bivariate...

Descripción completa

Detalles Bibliográficos
Autores: Rosa, G. J. M. [UNESP], Gianola, D.
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2001
País:Brasil
Institución:Universidade Estadual Paulista (UNESP)
Repositorio:Repositório Institucional da UNESP
Idioma:inglés
OAI Identifier:oai:repositorio.unesp.br:11449/66708
Acceso en línea:http://hdl.handle.net/11449/66708
Access Level:acceso abierto
Palabra clave:Bayesian inference
Bootstrap
EM algorithm
Maximum likelihood
Monte Carlo
Reference prior
Rejection sampling
Descripción
Sumario:The study of the association between two random variables that have a joint normal distribution is of interest in applied statistics; for example, in statistical genetics. This article, targeted to applied statisticians, addresses inferences about the coefficient of correlation (ρ) in the bivariate normal and standard bivariate normal distributions using likelihood, frequentist, and Baycsian perspectives. Some results are surprising. For instance, the maximum likelihood estimator and the posterior distribution of ρ in the standard bivariate normal distribution do not follow directly from results for a general bivariate normal distribution. An example employing bootstrap and rejection sampling procedures is used to illustrate some of the peculiarities.