Bayes prediction of the regression coefficient in a finite population using balanced loss function

摘要

In this paper, we derive Bayes predictor of the finite population regression coefficient of a linear regression normal superpopulation model under Zellner's(1994) balanced loss function. Loss robustness of the Bayes predictor is investigated in terms of relative losses and relative savings loss. The results are illustrated for location superpopulation model and also for a model with diagonal covariance error matrix. Some of the well known classical predictors are found to be a limiting case of the derived Bayes predictor. Sensitivity of Bayes predictor to a possible misspecification of the loss function is examined for a hypothetical population.