A Geometrical Explanation of Stein Shrinkage

TL;DR

This paper provides a geometric perspective on Stein's inadmissibility result for the multivariate normal mean estimator, using symmetry and visualization to clarify the underlying reasons and extend the argument to higher dimensions.

Contribution

It introduces a geometric framework that explains Stein's inadmissibility and extends the reasoning to higher dimensions, enhancing understanding of shrinkage estimators.

Findings

Geometric visualization clarifies Stein's inadmissibility

Inadmissibility extends to dimensions p ≥ 3

Provides intuitive and computational insights into shrinkage estimation

Abstract

Shrinkage estimation has become a basic tool in the analysis of high-dimensional data. Historically and conceptually a key development toward this was the discovery of the inadmissibility of the usual estimator of a multivariate normal mean. This article develops a geometrical explanation for this inadmissibility. By exploiting the spherical symmetry of the problem it is possible to effectively conceptualize the multidimensional setting in a two-dimensional framework that can be easily plotted and geometrically analyzed. We begin with the heuristic explanation for inadmissibility that was given by Stein [In Proceedings of the Third Berkeley Symposium on Mathematical Statistics and Probability, 1954--1955, Vol. I (1956) 197--206, Univ. California Press]. Some geometric figures are included to make this reasoning more tangible. It is also explained why Stein's argument falls short of yielding a proof of inadmissibility, even when the dimension, $p$, is much larger than $p=3$. We then extend the geometric idea to yield increasingly persuasive arguments for inadmissibility when $p\geq3$, albeit at the cost of increased geometric and computational detail.