Modelos Para Datos Longitudinales: Una propuesta Bayesiana
Cargando...
Autores
Castillo Carreño, Edwin Javier
Director
Tipo de contenido
Document language:
Español
Fecha
Título de la revista
ISSN de la revista
Título del volumen
Documentos PDF
Resumen
En este trabajo se presenta una propuesta Bayesiana para la estimación conjunta de estructuras de media y matriz de varianza y covarianza para modelos con datos longitudinales donde se suponen estructuras de covarianza SC, AR(1), ARMA(1; 1) y algunos casos de estructuras de antedependencia no estructurada de orden 1 ADS(1). Se generaliza el modelo para datos longitudinales con estructuras de antedependencia estructurada más general que la propuesta en Zimmerman, Núñez-antón and El-Barmi (1998). Se validan las distintas propuestas mediante simulaciones y aplicaciones a datos sugeridos en la literatura.
Abstract: In this work, is presented a Bayesian proposal for the joint estimation of structures of mean and variance and covariance matrix for models with longitudinal data where we assume structures of covariance SC, AR(1), ARMA(1:1) and some cases of structures of antedependence unstructured of order 1 ADS(1). Is generalized a model for longitudinal data with structures of structured antedependence more general than the one proposed in citeasnoun Zimmerman2. The different proposals are validated through simulations and applications to data suggested in the literature.
Abstract: In this work, is presented a Bayesian proposal for the joint estimation of structures of mean and variance and covariance matrix for models with longitudinal data where we assume structures of covariance SC, AR(1), ARMA(1:1) and some cases of structures of antedependence unstructured of order 1 ADS(1). Is generalized a model for longitudinal data with structures of structured antedependence more general than the one proposed in citeasnoun Zimmerman2. The different proposals are validated through simulations and applications to data suggested in the literature.