Good overview of Stein's Paradox in stats. Ex: Estimates for >=3 measurements (e.g. average height of EU, US, UK citizens) taken independently using MLE are beat (in expectation), when taken collectively, by a method appending a shrinking parameter. h
@johncarlosbaez @jjcherian @JustinRising I don't have a good source as I'm not an expert. I read somewhere that the best version to date was discovered in 1994, and somehow they proved it was not admissible. Unrelated paper, you might find interesting ht
This is news to me. But googling reveals that @fitelson has written about this, with Olav Vassend and Elliott Sober. https://t.co/5p76qLmMYR
Stein’s Paradox #SteinsParadox Charles Stein discovered a paradox in 1955 that many statisticians think is of fundamental importance, they explore its #philosophical implications here: https://t.co/vw20j9zXSy
The Philosophical Significance of Stein's Paradox https://t.co/JAFn6SR6on
New paper with Benjamin Vassend and Elliott Sober on Stein's Paradox. https://t.co/S7ohHWGu2u