Bayesian analysis of data in the health, pysical and social sciences has been greatly facilitated in the last decade by advances in computing power and improved scope for estimation via iterative sampling method. Yet, the bayesian perspective which stresses the accumulation of knowledge about parameters in a synthesis of prior knowledge with the data at hand,has a longer history. Suppose interest lies in rates of two mutually exclusive events,such as death due to digestive disease or death due to any cause. A confounding problem in the estimation of these rates is that counts for the cause of death may be misclassified. This work derives closed-form Bayesian estimators for two complementary Poisson rate parameters under the Entropy loss function, using mutually exclusive counted data subject to misclassification.
|