EE3331

Chapter 4: Conditional Distribution Functions

The slides are based on the notes of Prof. So.

Page 1 Semester B 2023-2024

Conditional Distribution Functions and Expectation

(i) Understand probability models conditioned by event or

random variable

(ii) Able to compute the conditional distribution functions for

one or two random variables

(iii) Able to compute expected values for the conditional

distributions

(v) Able to apply conditional probability models to solve

problems

Page 2 Semester B 2023-2024

Distribution Conditioned by Event
The concept of conditional probability can be extended to the
cumulative distribution function (CDF), probability mass
function (PMF) and probability density function (PDF) as they
are probability measures.
For discrete random variable (RV) and event , following
(1.7), the conditional PMF of given is defined as:

(4.1)

The conditional CDF is:

(4.2)

Page 3 Semester B 2023-2024

For continuous RV , the conditional CDF is same as (4.2)
and conditional PDF is:

(4.3)

Example 4.1
A Web distributes instructional videos on bicycle repair. The
length of a video is a RV in min. with PMF:

Suppose the Web has two servers, one for videos shorter
than 5 min. and the other for videos with at least 5 min. What
is the PMF of video length in the second server?
Page 4 Semester B 2023-2024
Let be the event when the second server is chosen, which
corresponds to . Hence we have:

According to (4.1),

Example 4.2
Consider a pointer-spinning experiment such that the pointer
will stop at a point with uniform distribution on the
circumference whose length is 1. Find the conditional PDF of
the pointer position for spins in which the pointer stops on
the left side of the circle.
Page 5 Semester B 2023-2024
Let RV be the pointer position. It is clear that
with PDF .

Denote as the event when the pointer stops on the left side
of the circle. Without loss of generality, we may assume that
corresponds to . Hence we have:

Page 6 Semester B 2023-2024

Analogous to (4.1), the conditional PDF is

Example 4.3
Suppose the time in integer minutes you wait for a bus is a
discrete RV with PMF:

Suppose the bus has not arrived by the 8th minute. What is
the conditional PMF of your waiting time ?

Page 7 Semester B 2023-2024

Let be the event , and we get
Hence:

Analogous to (1.8), law of total probability for PMF/PDF is:

(4.4)

where events , , are pairwise disjoint.

Example 4.4
Let denote the number of additional years that a randomly
chosen 70-year-old man will live. If he has abnormal blood
pressure, denoted as event , then has geometric PMF
Page 8 Semester B 2023-2024
with . If he has normal blood pressure, denoted as
event , then has geometric PMF with . Find the
conditional PMFs and . If 40% of all 70-year-old
men have abnormal blood pressure, what is the PMF of ?
According to (2.6), we have 2 conditional PMFs:

Since and , using (4.4) yields:

Page 9 Semester B 2023-2024

Example 4.5
Let RV be a received signal. When bits “0” and “1” are
transmitted, corresponding to events and , the received
values are and , respectively. Given
that bits “0” and “1” are equally likely to be sent, what is the
PDF of ?
It is clear that and are pairwise disjoint, and
. Furthermore, we have:

Applying (4.4) yields:

Page 10 Semester B 2023-2024

Conditional Expected Value given Event
Expectation operations can also be extended to the
conditional probability functions.
Given that the event has occurred, the conditional
expectation of can be extended from (2.19) and (2.20) as:

(4.5)

(4.6)

Consider events , , with and

, , we have:
(4.7)

Page 11 Semester B 2023-2024

Without loss of generality, for discrete RV, the proof is
obtained using (4.4) and (4.5):

Example 4.6
Compute of the geometric RV with parameter using
conditional expectation.

Assign as the event of getting success in the first trial and

recall that the geometric RV is the number of trials until the
first success occurs. Applying (4.7) yields:
Page 12 Semester B 2023-2024
Note that because when the first trial gives a
success, must be 1. While contains 1 due to the
first failure and accounts for the expected value
starting from the second trial because of independence.
As in (2.24)-(2.25), (4.5)-(4.6) can be generalized to a
function of RV , i.e., :
(4.8)

(4.9)

Page 13 Semester B 2023-2024

Analogous to (2.23), the conditional variance of given is:

(4.10)

Example 4.7
Find the conditional expected value and conditional variance
for the video length in the second server in Example 4.1.

(min.)

(min. )

(min. )
Page 14 Semester B 2023-2024
Conditional Joint Distributions given Event
For RVs and , and an event , the conditional joint
PMF/PDF of and given is:

(4.11)

Similarly, for , (4.8)-(4.9) can be extended to

(4.12)

(4.13)

The conditional variance extends (4.10) straightforwardly.

Page 15 Semester B 2023-2024
Example 4.8
Discrete RVs and have the joint PMF as shown
below. Find the conditional joint PMF given the
event . Then compute the conditional
expected value and conditional variance of .

It is clear that . Hence

Page 16 Semester B 2023-2024

Using (4.12), we obtain:

Similarly,

Page 17 Semester B 2023-2024

Finally, we have:

Example 4.9
Suppose two RVs and have joint PDF:

Find the conditional joint PDF given the event

. Then find the conditional expected value of
given .

Page 18 Semester B 2023-2024

Using (4.11), we obtain:

Applying (4.13) yields:

Page 19 Semester B 2023-2024

Distribution and Expectation Conditioned by Random Variable
Apart from conditioning by event, probability models can also
be conditioned by RV.

The definitions basically follow those of distribution

conditioned by event.
The conditional PMF/PDF of RV given RV is:

(4.14)

Similarly, the conditional PMF/PDF of given is

(4.15)
Page 20 Semester B 2023-2024
From (4.14) and (4.15), we have:

(4.16)

When and are independent, (4.14) becomes:

(4.17)

Similar to (4.4), for discrete RVs, we have:

(4.18)

Or more generally:
(4.19)

Page 21 Semester B 2023-2024

Similar to (4.5)-(4.10), the conditional expected values are:
(4.20)

(4.21)

(4.22)

(4.23)

(4.24)

(4.25)
Page 22 Semester B 2023-2024
Note that conditioning for continuous RVs is treated in a
similar manner, e.g.,

Example 4.10
Given the joint PDF of two RVs and :

Determine the value of and .

Applying (3.8), we have

Page 23 Semester B 2023-2024

To find , we utilize:

Hence:

Finally, applying (4.15) yields:

for .

Page 24 Semester B 2023-2024

Example 4.11
Suppose the joint PDF of RVs and is:

Find the conditional PDF for and conditional

PDF for . Then find the conditional expected
value .
For , we compute:

Page 25 Semester B 2023-2024

Applying (4.15) yields:

That is, for a given , . Similarly, for :

That is, for a given , .

Page 26 Semester B 2023-2024
Then we use (4.21):

Note that is a function of . If is not specified,

we can write:

which is also a RV, namely, a function of .

Page 27 Semester B 2023-2024

As is a function of , the expected value of
applies on only, and the result is:

(4.26)

This two-step process is known as iterated expectation.

Assuming continuous RVs, the proof can be derived as:

Page 28 Semester B 2023-2024

Iterated expectation can be generalized to :

Example 4.12
Apply iterated expectation to find in Example 4.11.

Recalling the result of , we apply (4.26) to obtain:

because

Page 29 Semester B 2023-2024

This aligns with

Example 4.13
Consider tossing a coin twice and the probability that the
outcome is head (H) is . Let and be the number of H,
either 0 or 1, in the first and second trials. Assign
and . Find , and .

Denote tail as T. We list out the possible outcomes first.

Outcome Probability
HH
HT
TH
TT
Page 30 Semester B 2023-2024
We can then construct the joint PMF of and :

Using (4.14), we have:

Page 31 Semester B 2023-2024

Page 32 Semester B 2023-2024
References:
1. R. D. Yates and D. J. Goodman, Probability and Stochastic
Processes: A Friendly Introduction for Electrical and Computer
Engineer, Wiley, 2014
2. S. Kay, Intuitive Probability and Random Processes using
MATLAB, Springer, 2006
3. H. Pishro-Nik, Introduction to Probability, Statistics, and
Random Processes, Kappa Research LLC, 2014
https://www.probabilitycourse.com

Page 33 Semester B 2023-2024