Page - 76 - in Differential Geometrical Theory of Statistics
Image of the Page - 76 -
Text of the Page - 76 -
Entropy2016,18, 386
Wewrite f the linear representationofLiealgebragofG,deïŹnedby fandq therestriction togof
thedifferential toq (f andq thedifferentialof fandq respectively),Koszulhasprovedthat:
f(X)q(Y)â f(Y)q(X)= q([X,Y]) âX,Yâ g
with f : gâ gl(E)andq : g âE (103)
wheregl(E) thesetofall linearendomorphismsofE, theLiealgebraofGL(E).
Usingthecomputation,
q(AdsY)= dq(s ·etY ·sâ1)
dt âŁâŁâŁâŁ
t=0 = f(s)f(Y)q(sâ1)+f(s)q(Y) (104)
Wecanobtain:
q([X,Y])= dq(AdetXY)
dt âŁâŁâŁâŁ
t=0 = f(X)q(Y)q(e)+f(e)f(Y)(âq(X))+ f(X)q(Y) (105)
where e is theunitelement inG. Since f(e) is the identitymappingandq(e)=0,wehave theequality:
f(X)q(Y)â f(Y)q(X)= q([X,Y]) .
Apair (f,q)ofa linearrepresentation f ofaLiealgebragonEanda linearmappingq fromg toE
isanafïŹnerepresentationofgonE, if it satisïŹes f(X)q(Y)â f(Y)q(X)= q([X,Y]) .
Conversely, if we assume that g admits an afïŹne representation (f,q) on E, using an afïŹne
coordinate system { x1,...,xn }
on E, we can express an afïŹnemapping v â f(X)v+q(Y) by an
(n+1)Ă(n+1)matrixrepresentation:
af f(X)= [
f(X) q(X)
0 0 ]
(106)
where f(X) isanĂnmatrixandq(X) isan rowvector.
X â af f(X) isan injectiveLiealgebrahomomorphismfromg in theLiealgebraofall (n+1)Ă
(n+1)matrices,gl(n+1,R): âŁâŁâŁâŁâŁ gâ
gl(n+1,R)X
â af f(X) (107)
Ifwedenotegaf f = af f(g),wewriteGaf f the linearLiesubgroupofGL(n+1,R)generatedby
gaf f . Anelementof sâGaf f isexpressedby:
Af f(s)= [
f(s) q(s)
0 1 ]
(108)
LetMaf f be the orbit ofGaf f through the origin o, thenMaf f = q(Gaf f) = Gaf f/Kaf f where
Kaf f = {
sâGaf f/q(s)=0 }
=Ker(q).
Example.LetΩbeaconvexdomain inRn containingnocompletestraight lines,wedeïŹneaconvex
coneV(Ω) inRn+1 = RnĂRbyV(Ω) = {(λx,x)âRnĂR/xâΩ,λâR+}. Then there exists an
afïŹneembedding:
: xâΩ â [
x
1 ]
âV(Ω) (109)
76
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