Seite - 219 - in Differential Geometrical Theory of Statistics
Bild der Seite - 219 -
Text der Seite - 219 -
Entropy2016,18, 433
Tosimlplifyweuse the followingnotation.
qpj = qMj.
In thecategoryFB(Î,Î)weconsideran isomorphism
[E1ĂM1]â (e,x)â [Κ(e),Ï(x)]âE2ĂM2.
(1) LetÏâbethedifferentialofÏ. ForââLC(M1) the imageÏâ(â)âLC(M2) isdeïŹnedby
[Ïâ(â)]XâYâ=Ïâ[âÏâ1â (Xâ)Ï
â1â (Yâ)]
forallvectorïŹeldsXâ,Yâ âX(M2).
(2) It is clear that thedatum [E1,Ï,M1,D1,p2âŠÎš] isanobjectof thecategoryGM(X,Ω). Thenfor
vectorïŹeldsX,Y inM1wecalculate (atX,Y) therighthandmemberof the followingequality
[qp2âŠÎš(â)]=â2[log(p2âŠÎš)].
Direct calculationsyield
â2[log(p2âŠÎš)](X,Y)=X · [Y · log(p2âŠÎš)]ââXY · log(p2âŠÎš)
=X · [Y · log(p2)âŠÎš]ââXY · [log(p2)âŠÎš]
=Ïâ(X) · [Ïâ(Y) · log(p2)]âÏâ(âXY) · log(p2)
= [Ïâ(â)2log(p2)](Ïâ(X),Ïâ(Y)).
Thus forallââLC(M1)wehave
q[p2âŠÎš](â)= qp2(Ïâ(â)).
Wesummarize thecalculations just carriedoutas it follows
Lemma8. Keeping thenotationwe justusednamely p2 andΚĂÏwehave the followingequality
q[p2âŠÎš] = qp2âŠÏâ
Weare inpositionto face theproblemofmoduli space in thecategoryGM(Î,Ω).
Theorem20. Weconsider twom-dimensional statisticalmodels
Mj=[Ej,Ïj,Mj,Dj,pj], j :=1,2.
In the categoryFB(Î,Î) letΚĂÏbean isomorphismof [E1,Ï1,M1,D1]onto [E2,Ï2,M2,D2].
The followingassertionsare equivalent.
(1) qp2âŠÏâ= qp1,
(2) p2âŠÎš= p1.
Demonstration.
Thedemonstration isbasedonLemmas7and8.
219
Differential Geometrical Theory of Statistics
- Titel
- Differential Geometrical Theory of Statistics
- Autoren
- Frédéric Barbaresco
- Frank Nielsen
- Herausgeber
- MDPI
- Ort
- Basel
- Datum
- 2017
- Sprache
- englisch
- Lizenz
- CC BY-NC-ND 4.0
- ISBN
- 978-3-03842-425-3
- Abmessungen
- 17.0 x 24.4 cm
- Seiten
- 476
- Schlagwörter
- Entropy, Coding Theory, Maximum entropy, Information geometry, Computational Information Geometry, Hessian Geometry, Divergence Geometry, Information topology, Cohomology, Shape Space, Statistical physics, Thermodynamics
- Kategorien
- Naturwissenschaften Physik