Bayes and robust Bayes prediction with an application to a rainfall prediction problem

In this paper, we consider the prediction problem of a future observation in a family of scale parameter models under a class of precautionary prediction loss function in the context of Bayes and robust Bayes methodology. Under three members of the precautionary prediction loss functions, which are suitable members when considering scale invariant models, we deal with Bayes, Minimax, Posterior Regret Gamma Minimax and Least Sensitive prediction of future observation YY based on currently observed data XX. We also conduct a simulation study and a data analysis to illustrate the practical utility of the prediction procedure and to compare performance of the proposed predictors, as well.