Scribd
Upload
27 views

Lecture #13: Conditional Expectation: Y - X X (Y X)

1) The document discusses a lecture on conditional expectation. It provides an example where X is the time of a radioactive decay and Y is the recorded time including a delay. The prior for X is exponential and the conditional of Y given X is also given. 2) Another example is presented where a stick of length 1 is broken randomly into two pieces. The expected value and variance of the second piece is sought. It is shown that the conditional expectation of Y given X=x is x/2. 3) The key concept discussed is that the conditional expectation of Y given X=x depends on the value of x, and can be written as a function of x.

Uploaded by

mohamed
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
27 views

Lecture #13: Conditional Expectation: Y - X X (Y X)

1) The document discusses a lecture on conditional expectation. It provides an example where X is the time of a radioactive decay and Y is the recorded time including a delay. The prior for X is exponential and the conditional of Y given X is also given. 2) Another example is presented where a stick of length 1 is broken randomly into two pieces. The expected value and variance of the second piece is sought. It is shown that the conditional expectation of Y given X=x is x/2. 3) The key concept discussed is that the conditional expectation of Y given X=x depends on the value of x, and can be written as a function of x.

Uploaded by

mohamed
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Statistics 351 (Fall 2015)

Prof. Michael Kozdron

October 7, 2015

Lecture #13: Conditional Expectation


Reference. 2.2 pages 3337
Example. Suppose that X is the time before the first occurrence of a radioactive decay
which is measured by an instrument. However, there is a delay built into the mechanism,
and the decay is recorded as having taken place at some time Y > X. (That is, although
we observe Y = y as our time of decay, the first occurrence of the decay actually takes place
at X = x. Because of the delay in the mechanism we know that y > x.) Assume that the
prior distribution for X is Exp(1) and that the conditional distribution of Y given X = x is
fY |X=x (y) = e
where

(y x)

0 < x < y < 1,

> 0 is a known constant. Determine the posterior density function fX|Y =y (x).

Solution. By definition,
fX|Y =y (x) =

fY |X=x (y)fX (x)


fX,Y (x, y)
=
.
fY (y)
fY (y)

We are told that


fY |X=x (y) = e

(y x)

and
fX (x) = e x ,
We now compute
Z 1
Z
fY (y) =
fX,Y (x, y) dx =
1

1
1

0 < x < y < 1,

0 < x < 1.
Z

fY |X=x (y)fX (x) dx =

= e
= e
=

y
0

e
Z

(y x)

(y x)

e
1

[1

e x
e (1

)y ]

(1
1

)e(
e(

(1

dx

)x

dx

1
1
y

for 0 < y < 1, and so we conclude


fX|Y =y (x) =

x=y

(1

)x
x=0

(1

)y

1)x
1)y

for 0 < x < y.


Example. A stick of length 1 is broken at a random point (uniformly over the stick). The
remaining piece is broken once more. Find the expected value and variance of the length of
the remaining piece.
131

Solution. Let X 2 U (0, 1) denote the position of the first random break, and let Y 2
U (0, X) denote the position of the second random break.

The interpretation of the distribution of Y is that given X = x, the random variable Y 2


U (0, x), i.e., Y |X = x 2 U (0, x) with X 2 U (0, 1), so that
(
1/x, for 0 < y < x,
fY |X=x (y) =
0,
otherwise.
Intuitively,
E(Y |X = x) =

x
and
2

Var(Y |X = x) =

x2
.
12

Formally, we have the following definition.


Definition. If (X, Y )0 is jointly distributed, then the conditional expectation of Y given
X = x is
8X
>
y pY |X=x (y),
in the discrete case, and
>
<
E(Y |X = x) = Zy 1
>
>
y fY |X=x (y) dy, in the continuous case.
:
1

provided that the sum or integral converges absolutely.

Example (continued). We can now verify that E(Y |X = x) = x/2. This follows since
Z 1
Z x
1
x
E(Y |X = x) =
yfY |X=x (y) dy =
y dy =
x
2
1
0
as expected.
Note. The conditional expectation of Y given X = x depends on the value of x. That is,
E(Y |X = x) is a function of x, say
E(Y |X = x) = h(x).

132

You might also like