Information projections revisited

被引：110

作者：

Csiszár, I

Matús, F

机构：

[1] Hungarian Acad Sci, Inst Math, H-1364 Budapest, Hungary

[2] Acad Sci Czech Republ, Inst Informat Theory & Automat, CR-18208 Prague, Czech Republic

来源：

IEEE TRANSACTIONS ON INFORMATION THEORY | 2003年 / 49卷 / 06期

关键词：

convex core; exponential family; I-projection; Kullback-Leibler divergence; maximum likelihood (ML); Pythagorean identity;

D O I：

10.1109/TIT.2003.810633

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 [计算机科学与技术];

摘要：

The goal of this paper is to complete results available about I-projections, reverse I-projections, and their generalized versions, with focus on linear and exponential families. Pythagorean-like identities and inequalities are revisited and generalized, and generalized maximum-likelihood (ML) estimates for exponential families are introduced. The main tool is a new concept of extension of exponential families, based on our earlier results on convex cores of measures.

引用

页码：1474 / 1490

页数：17

共 23 条

[1]

Information geometry on hierarchy of probability distributions [J].