In statistics, the matrix variate beta distribution is a generalization of the beta distribution. If
U
p x p
a,b>(p-1)/2
U\simBp\left(a,b\right)
I\left(a,b\right) | |
B | |
p |
U
-1 | |
\left\{\beta | |
p\left(a,b\right)\right\} |
\det\left(U\right)a-(p+1)/2
b-(p+1)/2 | |
\det\left(I | |
p-U\right) |
.
Here
\betap\left(a,b\right)
\beta | ||||
|
where
\Gammap\left(a\right)
\Gammap\left(a\right)=\pip(p-1)/4
p\Gamma\left(a-(i-1)/2\right). | |
\prod | |
i=1 |
If
U\simBp(a,b)
X=U-1
1 | |
\betap\left(a,b\right) |
\det(X)-(a+b)
b-(p+1)/2 | |
\det\left(X-I | |
p\right) |
X>Ip
a,b>(p-1)/2
If
U\simBp(a,b)
H
p x p
HUHT\simB(a,b).
Also, if
H
p x p
U
HUHT\simBp(a,b)
H
If
A
q x p
q\leqp
q
AUAT
AUAT\sim
T,0\right) | |
GB | |
q\left(a,b;AA |
If
U\simBp\left(a,b\right)
U
U=\begin{bmatrix} U11&U12\\ U21&U22\end{bmatrix}
where
U11
p1 x p1
U22
p2 x p2
U22 ⋅
U22-U21{U11
U11
U22 ⋅
U11\sim
B | |
p1 |
\left(a,b\right)
U22 ⋅ \sim
B | |
p2 |
\left(a-p1/2,b\right)
U21\midU11,U22 ⋅
U21\midU11,U22 ⋅ \sim
IT | |
p2,p1 |
\left(2b-p+1,0,I | |
p2 |
-U22 ⋅ ,U11
(I | |
p1 |
-U11)\right).
Mitra proves the following theorem which illustrates a useful property of the matrix variate beta distribution. Suppose
S1,S2
p x p
S1\simWp(n1,\Sigma),S2\simWp(n2,\Sigma)
\Sigma
n1+n2\geqp
U=S-1/2
-1/2 | |
S | |
1\left(S |
\right)T,
where
S=S1+S2
U
Bp(n1/2,n2/2)
U
\Sigma