An Entropy-Energy Identity for Predictive Kullback-Leibler Regret in Infinitely Divisible Location Models

Kōsaku Takanashi, Kenichiro McAlinn · 2026

Read An Entropy-Energy Identity for Predictive Kullback-Leibler Regret in Infinitely Divisible Location Models by Kōsaku Takanashi, Kenichiro McAlinn (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-27253.txt.

Description

Kosaku Takanashi and Kenichiro McAlinn study how well competing prediction rules perform for a class of location models. They prove an exact identity linking Kullback-Leibler prediction regret to the energy of a related Markov process. That link turns regret comparisons into a question about recurrence, and yields tail conditions under which a Bayes prediction rule cannot be beaten.

Catalog record

CREATORKōsaku Takanashi, Kenichiro McAlinn
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE83440
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.27253

Browse the full library · About the Tower of Babel