Robust Bayesian estimator in a normal model with uncertain hierarchical priors

In this paper, robust Bayesian estimators of error variance in a normal linear model with uncertain hierarchical prior information are investigated. The posterior regret gamma minimax estimator, the least sensitive estimator and conditional gamma minimax estimator for error variance are obtained under two different classes of priors, respectively. A simulation study is used to compare the performance of the proposed estimators. A real data example is also given to illustrate the results.