|
BAYESIAN INFERENCE FOR THE PARAMETER OF THE POWER DISTRIBUTIONKeywords: Bayes Factor , Bayes Posterior Risk , Elicitation , Hyperparameter Abstract: This study provides Bayesian analysis of the power model using two informative(gamma and Rayleigh) priors and two non-informative (Jeffreys and uniform) priors. The priorpredictive distribution is used to elicit the values of the hyperparameters of the prior distribution.The priors are compared using Bayes point and interval estimates, posterior variances,coefficients of skewness and coefficients of kurtosis. Bayes factors and Bayes posterior risks arealso used for the comparison of informative and non-informative priors.
|