On statistics which are almost sufficient from the viewpoint of the Fisher metrics

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yamaguchi, Kaori, Nozawa, Hiraku
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910458920828928
author Yamaguchi, Kaori
Nozawa, Hiraku
author_facet Yamaguchi, Kaori
Nozawa, Hiraku
contents A statistic on a statistical model is sufficient if it has no information loss, namely, the Fisher metric of the induced model coincides with that of the original model due to Kullback and Ay-Jost-Lê-Schwachhöfer. We introduce a quantitatively weak version of sufficient statistics such that the Fisher metric of the induced model is bi-Lipschitz equivalent to that of the original model. We characterize such statistics in terms of the conditional probability or by the existence of a certain decomposition of the density function in a way similar to characterizations of sufficient statistics due to Fisher-Neyman and Ay-Jost-Lê-Schwachhöfer.
format Preprint
id arxiv_https___arxiv_org_abs_2305_04199
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On statistics which are almost sufficient from the viewpoint of the Fisher metrics
Yamaguchi, Kaori
Nozawa, Hiraku
Statistics Theory
Differential Geometry
Probability
A statistic on a statistical model is sufficient if it has no information loss, namely, the Fisher metric of the induced model coincides with that of the original model due to Kullback and Ay-Jost-Lê-Schwachhöfer. We introduce a quantitatively weak version of sufficient statistics such that the Fisher metric of the induced model is bi-Lipschitz equivalent to that of the original model. We characterize such statistics in terms of the conditional probability or by the existence of a certain decomposition of the density function in a way similar to characterizations of sufficient statistics due to Fisher-Neyman and Ay-Jost-Lê-Schwachhöfer.
title On statistics which are almost sufficient from the viewpoint of the Fisher metrics
topic Statistics Theory
Differential Geometry
Probability
url https://arxiv.org/abs/2305.04199