7-4
Therefore, the desired distribution is
We d efine the Bayes estimatorof as the value that corresponds to the mean of the poste-
rior distribution
Sometimes, the mean of the posterior distribution of can be determined easily. As a
function of , is a probability density function and are just con-
stants. Because enters into only through if
as a function of is recognized as a well-known probability function, the posterior mean of
can be deduced from the well-known distribution without integration or even calculation of
EXAMPLE S7-2 Let X 1 , X 2 , p, Xnbe a random sample from the normal distribution with mean and variance
^2 , where is unknown and ^2 is known. Assume that the prior distribution for is normal
with mean 0 and variance ; that is
The joint probability distribution of the sample is
Thus, the joint probability distribution of the sample and is
Upon completing the square in the exponent
where hi(x 1 , p, xn, ^2 , 0 , ) is a function of the observed values, ^2 , 0 , and.
Now, because f(x 1 , p, xn) does not depend on ,
f 1 0 x 1 ,p, xn 2 e
(^11) 22 a^1
02
1
^2 nb^ c
(^2) a^1 ^2 n^2 ^0 ^20 x
^20 ^2 n bd^ h
31 x 1 ,p, xn,
(^2) ,
0 ,
2
02
^20 ^20
f 1 x 1 , x 2 ,p, xn, 2 e
(^11) 22 a^1
02
1
^2 nb^ c
(^2) a^1
(^2) n 2 0
^20 ^2 n^
(^)
x^20
^20 ^2 nb
d
2
(^) h 21 x 1 ,p, xn, ^2 , 0 , ^202
e
(^11) 22 ca^1
02
1
^2 nb^
(^2) 2 a^0
02
x
^2 nb^ d^ h 11 x 1 ,p, xn, ^2 , 0 , ^202
f 1 x 1 , x 2 ,p, xn, 2
1
12
^22 n^212
0
e^11 ^22311
(^20) n (^22) (^2) 12 0 (^20 2) a xi (^22) a x (^2) i (^2) (^20) (^204)
1
12
^22 n^2
e^11 ^2
(^221) ax (^2) i 2 axi n (^22)
f 1 x 1 , x 2 ,p, xn 0 2
1
12
^22 n^2
e^11 ^2
(^22) an
i 11 xi^2
2
f 1 2
1
12
0
e^1 ^02
(^2) 12 (^202)
1
12
^20
e^1
(^2) 2 0 (^202) 12 (^202)
^20
f 1 x 1 ,p, xn 2.
f 1 0 x 1 ,p, xn 2 f 1 x 1 ,p, xn, 2 f 1 x 1 ,p, xn, 2 ,
f 1 0 x 1 ,p, xn 2 x 1 ,p, xn
f 1 0 x 1 , x 2 ,p, xn 2.
f 1 0 x 1 , x 2 ,p, xn 2
f 1 x 1 , x 2 ,p, xn, 2
f 1 x 1 , x 2 ,p, xn 2
PQ220 6234F.CD(07) 5/15/02 4:31 PM Page 4 RK UL 6 RK UL 6:Desktop Folder:TEMP WORK:MONTGOMERY:REVISES UPLO D CH114 FIN L:Quark F