Answers are grounded in this paper and cite their sources.
Suppose we have k(≥2) normal populations with an unknown common mean μ and possibly different unknown variances σi2; i=1,2,…,k, respectively. The problem of estimation of μ with respect to a quadratic loss function is considered. Sufficient conditions for inadmissibility of estimators in certain classes of affine and location have been derived. Consequently, some complete class results are obtained. A detailed simulation study has been carried out in order to numerically compare the risk functions of some well-known estimators of common mean for the case k=3 and k=4. Finally, certain decisions have been taken regarding the use of these estimators.
No takes yet. Share an insight, caveat, or question.
Tripathy et al. (2015) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: