On divergence of finite measures and their applicability in statistics and information theory

0329681 - ÚTIA 2010 RIV GB eng J - Journal Article
Vajda, Igor - Stummer, W.
On divergence of finite measures and their applicability in statistics and information theory.
[O divergencích konečných měr a jejich využití ve statistice a teorii informace.]
Statistics. Roč. 44, č. 2 (2009), s. 169-187. ISSN 0233-1888. E-ISSN 1029-4910
R&D Projects: GA MŠMT(CZ) 1M0572; GA ČR(CZ) GA102/07/1131
Institutional research plan: CEZ:AV0Z10750506
Keywords : Local and global divergences of finite measures * Divergences of sigma-finite measures * Statistical censoring * Pinsker's inequality, Ornstein's distance * Differential power entropies
Subject RIV: BD - Theory of Information
Impact factor: 0.759, year: 2009
http://library.utia.cas.cz/separaty/2009/SI/vajda-on divergence of finite measures and their applicability in statistics and information theory.pdf

Family of divergences of finite and sigma-finite measures is introduced. Range of values, symmetry and decomposition into local and global components are obtained. Censoring is used to illustrate applications in statistics. Pinsker's inequality and Ornstein's distance of stationary random processes are among the applications in information theory.

Je navržena soustava divergencí konečných a sigma-konečných měr. Pro ně jsou prozkoumány obory hodnot, symetričnost a rozklady na lokalní a globální komponenty. Cencorování ilustruje užitečnost ve statistice. Pinskerova nerovnost a Ornsteinova vzdálenost stacinárních procesů ilustrují využití ve statistice.
Permanent Link: http://hdl.handle.net/11104/0175649