Geoscience Reference
In-Depth Information
So, X e Y independent
!
EXY
ðÞ
¼
EX
ðÞ
EY
ðÞ;
Cov X
ð
;
Y
Þ
¼
0
;
q
XY
¼
0
and
VX
ð
þ
Y
Þ
¼
VX
ðÞþ
VY
ðÞ:
A.6 Conditional Distributions
A more complete information on the dependence between random variables is
given by the conditional expectations.
For any p and any random variables X and Y de
∈
R for which p[X=x] > 0, the expected value of Y with respect to p conditional on
the event {s
ned on (S,
ʛ
) and any x
S | X(s) = x}, that means with respect to the probability p(|[X = x])
can be computed. It is denoted E(Y|X = x).
If X is a discrete random variable, a function Y
|X
with domain S and codomain
R that for each s with p[X = X(s)] > 0 assigns Y
|X
(s) = E(Y|X = X(s)) is a random
variable and its expected value E(EY
|X
) equals the expected value EY of Y.
More generally, for any A
∈
, the expected value of the restriction of Y
|X
to A
coincides with the expected value of the restriction of Y to A, that means,
EY
j
X
1
A
∈ ʛ
¼
EY1
A
ð
Þ:
Or extending yet a little more, for any random variable g(X, Y)
EgX
:
ð
ð
;
Y
Þ
Þ
¼
EgX
;
EY
j
X
nition of conditional expectation.
For any X and Y and any Z that is constant in any set of
This property is employed to extend the de
where X is constant and
Λ
satis
es E(g(X, Z)) = E(g(X, Y)), Z is a conditional expectation of Y given X.
This can be extended to conditioning on a vector of random variables. It is
enough to replace in the above formulation the random variable X by a vector X =
(X
1
,
…
,X
n
) of random variables X
1
,
…
,X
n
.
Replacing Y by 1
[Y
∈
A]
,
for each event A,
this de
nition of conditional
expectation can be used to de
ne a distribution of Y conditional on X = x, denoted
by p
Y
(|X = x) such that a conditional expectation of Y given X can be obtained
computing, for each x, the expectation of Y with respect to this distribution of Y
conditional on X = x.
In the discrete case, the distribution of Y conditional on X = x is given by
p
Y
A
½
j
X
¼
x
¼
p
ð
X
¼
x
;
Y
2
A
Þ=
pX
ð
¼
x
Þ:
For continuous random vectors, the conditional density of Y given X is, for each
real x, the density of the conditional distribution of Y given X = x, that is the real
Search WWH ::
Custom Search