Page - 219 - in Differential Geometrical Theory of Statistics
Image of the Page - 219 -
Text of the Page - 219 -
Entropy2016,18, 433
Tosimlplifyweuse the followingnotation.
qpj = qMj.
In thecategoryFB(Γ,Ξ)weconsideran isomorphism
[E1×M1]∈ (e,x)→ [Ψ(e),ψ(x)]∈E2×M2.
(1) Letψ∗bethedifferentialofψ. For∇∈LC(M1) the imageψ∗(∇)∈LC(M2) isdefinedby
[ψ∗(∇)]X∗Y∗=ψ∗[∇ψ−1∗ (X∗)ψ
−1∗ (Y∗)]
forallvectorfieldsX∗,Y∗ ∈X(M2).
(2) It is clear that thedatum [E1,π,M1,D1,p2◦Ψ] isanobjectof thecategoryGM(X,Ω). Thenfor
vectorfieldsX,Y inM1wecalculate (atX,Y) therighthandmemberof the followingequality
[qp2◦Ψ(∇)]=∇2[log(p2◦Ψ)].
Direct calculationsyield
∇2[log(p2◦Ψ)](X,Y)=X · [Y · log(p2◦Ψ)]−∇XY · log(p2◦Ψ)
=X · [Y · log(p2)◦Ψ]−∇XY · [log(p2)◦Ψ]
=ψ∗(X) · [ψ∗(Y) · log(p2)]−ψ∗(∇XY) · log(p2)
= [ψ∗(∇)2log(p2)](ψ∗(X),ψ∗(Y)).
Thus forall∇∈LC(M1)wehave
q[p2◦Ψ](∇)= qp2(ψ∗(∇)).
Wesummarize thecalculations just carriedoutas it follows
Lemma8. Keeping thenotationwe justusednamely p2 andΨ×ψwehave the followingequality
q[p2◦Ψ] = qp2◦ψ∗
Weare inpositionto face theproblemofmoduli space in thecategoryGM(Ξ,Ω).
Theorem20. Weconsider twom-dimensional statisticalmodels
Mj=[Ej,πj,Mj,Dj,pj], j :=1,2.
In the categoryFB(Γ,Ξ) letΨ×ψbean isomorphismof [E1,π1,M1,D1]onto [E2,π2,M2,D2].
The followingassertionsare equivalent.
(1) qp2◦ψ∗= qp1,
(2) p2◦Ψ= p1.
Demonstration.
Thedemonstration isbasedonLemmas7and8.
219
Differential Geometrical Theory of Statistics
- Title
- Differential Geometrical Theory of Statistics
- Authors
- Frédéric 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