Web-Books
in the Austria-Forum
Austria-Forum
Web-Books
Naturwissenschaften
Physik
Differential Geometrical Theory of Statistics
Page - 275 -
  • User
  • Version
    • full version
    • text only version
  • Language
    • Deutsch - German
    • English

Page - 275 - in Differential Geometrical Theory of Statistics

Image of the Page - 275 -

Image of the Page - 275 - in Differential Geometrical Theory of Statistics

Text of the Page - 275 -

Entropy2016,18, 407 3.GeneralizationofRĆ©nyiDivergence In this section, weprovide a generalization of RĆ©nyi divergence, which is given in terms of aĻ•-function. Thisgeneralizationalsodependsonaparameterα∈ [āˆ’1,1]; forα=±1, it isdefined as a limit. Supposing that theunderlying Ļ•-function is continuouslydifferentiable,we show that this limitexistsandresults in theĻ•-divergence [12]. Inwhat follows,allprobabilitydistributionsare assumedtohavepositivedensity. Inotherwords, theybelongto thecollection Pμ= { p∈L0 : ∫ T pdμ=1and p>0 } , where L0 is the space of all real-valued, measurable functions on T, with equality μ-a.e. (μ-almosteverywhere). TheRĆ©nyidivergenceoforderα∈ (āˆ’1,1)betweentwoprobabilitydistributions pandq inPμ is definedas D(α)R (p‖ q)= 4 α2āˆ’1 log (∫ T p 1āˆ’Ī± 2 q 1+α 2 dμ ) . (4) Forα=±1, theRĆ©nyidivergence isdefinedbytakinga limit: D(āˆ’1)R (p‖ q)= limĪ±ā†“āˆ’1D (α) R (p‖ q), (5) D(1)R (p‖ q)= limα↑1D (α) R (p‖ q). (6) Under some conditions, the limits in (5) and (6) are finite-valued, and converge to the Kullback–Leiblerdivergence. Inotherwords, D(āˆ’1)R (p‖ q)=D(1)R (q‖ p)=DKL(p‖ q)<āˆž, whereDKL(p‖ q)denotes theKullback–Leiblerdivergencebetween pandq,which isgivenby DKL(p‖ q)= ∫ T p log (p q ) dμ. TheseconditionsarestatedinProposition1,givenintheendof thissection, for thecase involving thegeneralizedRĆ©nyidivergence. TheRĆ©nyidivergence in its standardformisgivenby D(α)(p‖ q)= 1 1āˆ’Ī± log (∫ T pαq1āˆ’Ī±dμ ) , forα∈ (0,1). (7) Expression(4) is relatedto this formby D(α)R (p‖ q)= 2 1āˆ’Ī±D ((1āˆ’Ī±)/2)(p‖ q). Beyondthechangeofvariables,whichresults inα rangingin [āˆ’1,1], expressions (4)and(7)differ bythe factor2/(1āˆ’Ī±).Weoptedto insert the term2/(1āˆ’Ī±) so that somekindofsymmetrycouldbe maintainedwhenthe limitsĪ±ā†“āˆ’1andα↑1areconsidered. Inaddition, thegeometry inducedbythe version(4)conformswithAmari’snotation[5]. TheRĆ©nyi divergenceD(α)R (Ā· ‖ Ā·) canbedefined for every α ∈R. However, for α /∈ (āˆ’1,1), theexpression(4)maynotbefinite-valuedforevery pandq inPμ. Toavoidsometechnicalities,we justconsiderα∈ [āˆ’1,1]. 275
back to the  book Differential Geometrical Theory of Statistics"
Differential Geometrical Theory of Statistics
Title
Differential Geometrical Theory of Statistics
Authors
FrƩdƩric Barbaresco
Frank Nielsen
Editor
MDPI
Location
Basel
Date
2017
Language
English
License
CC BY-NC-ND 4.0
ISBN
978-3-03842-425-3
Size
17.0 x 24.4 cm
Pages
476
Keywords
Entropy, Coding Theory, Maximum entropy, Information geometry, Computational Information Geometry, Hessian Geometry, Divergence Geometry, Information topology, Cohomology, Shape Space, Statistical physics, Thermodynamics
Categories
Naturwissenschaften Physik
Web-Books
Library
Privacy
Imprint
Austria-Forum
Austria-Forum
Web-Books
Differential Geometrical Theory of Statistics