Gómez Villegas, Miguel Á. and Main Yaque, Paloma and Navarro, H. and Susi García, Rosario (2010) Sensitivity to hyperprior parameters in Gaussian Bayesian networks. [ Cuadernos de Trabajo de la Escuela Universitaria de Estadística; nº 03/201, ]

PDF
1MB 
Official URL: http://www.ucm.es/info/eue//pagina/cuadernos_trabajo/ct03_2010.pdf
Abstract
Our focus is on learning Gaussian Bayesian networks (GBNs) from data. In GBNs the multivariate normal joint distribution can be alternatively specified by the normal regression models of each variable given its parents in the DAG (directed acyclic graph). In the later representation the paramenters are the mean vector, the regression coefficients and the corresponding conditional variances. the problem of Bayesian learning in this context has been handled with different approximations, all of them concerning the use of different priors for the parameters considered we work with the most usual prior given by the normal/inverse gamma form. In this setting we are inteserested in evaluating the effect of prior hyperparameters choice on posterior distribution. The KullbackLeibler divergence measure is used as a tool to define local sensitivity comparing the prior and posterior deviations. This method can be useful to decide the values to be chosen for the hyperparameters.
Item Type:  Working Paper or Technical Report 

Uncontrolled Keywords:  Gaussian Bayesian networks, KullbackLeibler divergence, Bayesian linear regression 
Subjects:  Sciences > Mathematics > Mathematical statistics Sciences > Statistics > Operations research 
Series Name:  Cuadernos de Trabajo de la Escuela Universitaria de Estadística 
Volume:  
Number:  03/201 
ID Code:  10941 
Deposited On:  30 Jun 2010 10:26 
Last Modified:  14 Mar 2016 11:31 
Repository Staff Only: item control page