Page - 240 - in Differential Geometrical Theory of Statistics
Image of the Page - 240 -
Text of the Page - 240 -
Entropy2016,18, 383
wherevk∈Rqk isavectorcorrespondingto theWk-componentofξ. Ifξ∈P∗V,wesee from(19) that
ψk(ξ) ispositivedefinite. In thiscase,wehave:
φk(ξ)= (
1 tvkψk(ξ)−1
Iqk )(
ξkk− tvkψk(ξ)−1vk
ψk(ξ) )(
1
ψk(ξ) −1vk Iqk )
, (21)
so thatwegetageneralizationof (18), that is,
(φk(ξ) −1)11=(ξkk− tvkψk(ξ)−1vk)−1. (22)
Ontheotherhand, ifqk=0, thenφk(ξ)−1= ξ−1kk .
Weremarkthatψ1(ξ)=ψ(ξ′), andthatsomepartof theargumentabove isparallel to theproof
ofTheorem1.
Theorem2. TheKoszul–Vinbergcharacteristic functionϕV ofP∗V isgivenby the following formula:
ϕV(ξ)=CV r
∏
k=1 (
φk(ξ) −1)1+qk/2
11 ∏
qk>0 (detψk(ξ)) −1/2 (ξ∈P∗V), (23)
whereCV :=(2π)(N−r)/2∏rk=1Γ(1+ qk
2 )andN :=dimZV.
Proof. Weshall showthe statementby inductionon the rankas in theproof ofTheorem1. Then,
it suffices toshowthat:
ϕV(ξ)=(2π)q1/2Γ(1+ q1
2 )(φ1(ξ) −1)1+q1/211 (detψ1(ξ)) −sgn(q1)/2ϕV′(ξ′) (24)
forξ∈P∗V as in (10),where (detψ1(ξ))−sgn(q1)/2 is interpretedas:
(detψ1(ξ))−sgn(q1)/2 := {
1 (q1=0),
(detψ1(ξ))−1/2 (q1>0).
Whenq1=0,wehave:
ϕV(ξ)= ∫ ∞
0 ∫
PV′ e−x11ξ11e−(x ′|ξ′)dx11dx′
= ξ−111 ϕV′(ξ ′),
whichmeans (24).
When q1> 0, theEuclideanmeasure dx equals 2q1/2x q1
11dx11du˜dx˜ ′ by the changeof variables
in (6). Indeed, thecoefficient2q1/2 comes fromthenormalizationof the innerproductonW Rq1
regardedasasubspaceofZV. Then,wehaveby(15):
ϕV(ξ)= ∫ ∞
0 ∫
R q1 ∫
PV′ e−x11(ξ11− tvψ(ξ′)−1v)e−x11 t(u˜+ψ(ξ′)−1v)ψ(ξ′)(u˜+ψ(ξ′)−1v)e−(x˜ ′|ξ′)
×2q1/2xq111dx11du˜dx˜′.
By theGaussian integral
formula,wehave:∫
R q1 e−x11 t(u˜+ψ(ξ′)−1v)ψ(ξ′)(u˜+ψ(ξ′)−1v)du˜=πq1/2x−q1/211 (detψ(ξ ′))−1/2.
240
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