Page - 88 - in Differential Geometrical Theory of Statistics
Image of the Page - 88 -
Text of the Page - 88 -
Entropy2016,18, 386
Figure11.AffineLiegroupactionformultivariateGaussian law.
Consideringthecurveγ(t)anditsderivative .
γ(t):
γ(t)= [
R1/2(t) m(t)
0 1 ]
and .
γ(t)= [ .
R 1/2
(t) .
m(t)
0 0 ]
(185)
Wecanconsider thecurvewith thepointγ(0)movedat the identityelementonthe leftoronthe
right. Then, the tangentplanat identityelementprovides theLiealgebra:
ΓL(t)= LM−1 (γ(t))= [
R−1/2R1/2(t) R−1/2(m(t)−m)
0 1 ]
(186)
.
ΓL(t) ∣∣∣
t=0 = [
R−1/2 .
R 1/2
(0) R−1/2 .m(0)
0 1 ]
= ddt (LM−1(γ(t))) ∣∣∣
t=0 = dLM−1 .
γ(0)= dLM−1 .
M (187)
Liealgebraontherightandonthe left is thedefinedby:
dLM−1 :TM(G)→ gL
.
M →ΩL= dLM−1 .
M=M−1 .
M= [
R−1/2 .
R 1/2 R−1/2 .m
0 0 ]
(188)
dRM−1 :TM(G)→ gR
.
M →ΩR= dRM−1 .
M= .
MM−1= [
R−1/2 .
R 1/2 .
m−R−1/2 .R1/2 .m
0 0 ]
(189)
Wecan thenobserve thevelocities in twodifferentways, either byplacing in afixedoutside
frame,eitherbyputting inplaceof theelement in theprocessofmovingbyplacing in thereference
frameof theelement.[
X(t)
1 ]
=M [
x
1 ]
⇒ [ .
X(t)
0 ]
=ΩR [
X(t)
1 ]
withxfixed (190)
[
x(t)
1 ]
=M−1 [
X
1 ]
⇒ [ .
x(t)
0 ]
=−ΩL [
X
1 ]
withXfixed (191)
In the following,wewill complete the global viewby the operatorswhichwill allow to link
algebra (fromthe leftor theright)betweenthemandalsoconnect to theirdual.Wewillfirst consider
88
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