On shrinkage estimation for balanced loss functions (Q2293387)

In the paper under review the estimation of a multivariate mean of an observable random vector \(X\) is studied under two natural modifications of balanced loss functions. The first of these modifications was considered earlier by [\textit{M. Jafari Jozani} et al., Stat. Pap. 53, No. 1, 51--60 (2012; Zbl 1241.62004)], where Bayesian estimators as well as other types of posterior risk analysis were provided. However, for both losses considered in the present paper, there seems to be no significant known finding for frequentist risk analysis (such as the earlier results for balanced squared error loss). The objective of this paper is to try to fill such gaps. To achieve this, the authors focus on the multivariate normal case for \(X\), as well as on a scale mixture of normals as defined in [\textit{A. C. Brandwein} et al., Ann. Inst. Stat. Math. 45, No. 3, 551--565 (1993; Zbl 0811.62058)]. Under certain conditions, shrinkage estimators of the mean that improve on the benchmark estimator \(X\) (as measured by the frequentist risk associated with presented balanced loss functions) are provided.