We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ME
< prev | next >
new | recent | 1911

Change to browse by:

math
math.PR
math.ST
physics
physics.data-an
quant-ph
stat

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Mathematics > Statistics Theory

Title: Update of prior probabilities by minimal divergence

Authors: Jan Naudts
(Submitted on 1 Nov 2019 (v1), last revised 13 Nov 2021 (this version, v2))
Abstract: The present paper investigates the update of an empirical probability distribution with the results of a new set of observations. The optimal update is obtained by minimizing either the Hellinger distance or the quadratic Bregman divergence. The results obtained by the two methods differ. Updates with information about conditional probabilities are considered as well.
Comments: Small improvements
Subjects: Statistics Theory (math.ST); Probability (math.PR); Data Analysis, Statistics and Probability (physics.data-an); Quantum Physics (quant-ph); Methodology (stat.ME)
MSC classes: 62G05
Journal reference: Entropy 23, 1668 (2021)
DOI: 10.3390/e23121668
Cite as: arXiv:1911.00217 [math.ST]
  (or arXiv:1911.00217v2 [math.ST] for this version)

Submission history

From: Jan Naudts [view email]
[v1] Fri, 1 Nov 2019 06:12:15 GMT (9kb)
[v2] Sat, 13 Nov 2021 16:23:48 GMT (8kb)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.