Admissible and minimax estimation of the parameter of the selected Pareto population under squared log error loss function (Q2359160)

The paper under review deals with admissible and minimax estimation of two parameters associated with the selected Pareto population under squared log error loss function. The uniformly minimum risk unbiased (UMRU) estimators of these parameters are provided. In the case of selection from two populations, the author presents a sufficient condition for estimators of considered parameters to be minimax and finds that both UMRU and natural estimators of the parameters are minimax. He also identifies the class of linear admissible estimators for these parameters, which contain the natural estimator. Next, utilizing the Brewester-Ziedeck methodology [\textit{J. F. Brewster} and \textit{J. V. Zidek}, Ann. Stat. 2, 21--38 (1974; Zbl 0275.62006)], the current author finds a sufficient condition for inadmissibility of scale and permutation invariant estimators of the second parameter and shows that UMRU estimator of that parameter is inadmissible. Numerical comparison of the risk of discussed estimators is also presented. Finally, citing [\textit{Z. Mohammadi}, ``Estimation after selection under squared log error loss function'', in: Proceedings of the 12th Iranian statistical conference. 167--170 (2014)], the paper under review depicts results related to the first of the discussed parameters, which can be obtained for selected uniform population by arguments similar to the ones presented here.