Analyzing the Prevalence of Overweight and Obesity of Pakistani Females in Bayesian Paradigm

Abstract

This study analyzed the prevalence of over weight and obesity of Pakistani female in Bayesian paradigm. For the modeling of this data set Rayleigh-Rayleigh distribution (RRD) is used. The posterior distributions are evaluated using uniform, Jeffreys and exponential priors. These posterior distributions are not attained in closed form. Two approximation techniques, Lindley and Tierney-Kadane (T-K) are used to obtain the Bayes estimators under three different loss functions (squared error, Weighted and Precautionary loss functions). Monte Carlo simulation study and real life data about the prevalence of over weight and obesity of the Pakistani female is used to show the superiority of Bayes estimators over the maximum likelihood (ML) estimators. It is concluded that Bayes estimators under informative priors perform better than non informative priors due to minimum associated risks. It is also found that estimators obtained through Bayesian technique are better than most common ML method.