1

Unit Six
Probability Distributions

seid.belay@aastu.edu.et 19-May-25
2 Introduction
 Probability distributions are fundamental concepts in statistics. They are
used both on a theoretical level and a practical level.
 Used for
 Confidence interval computations
 Hypothesis testing
 Modeling of data
 Simulations …

seid.belay@aastu.edu.et 19-May-25
3 Random Variable (RV)
 Given random experiment with sample space S. Then a random variable is
a real valued function that maps elements of S to real number.
X
 S ՜ 𝑅 X is Called RV, X S = R, R is called the range of X

- Takes values in the set 0,1,2, …

Discrete RV - Example: Two teams A and B play a game
against each other home and away. A team
either wins(W), (D)draw or (L)loses a match.
 Random Variable - Let X = No of times Team A wins:
- Frequency Distribution of X: X 2 1 0
f 1 4 4

Continuous RV - Can take any value in an

interval / union of intervals
- Example: Waiting time to get
service in a service station
- 𝑇: 𝑡 ∈ (0, ∞)
seid.belay@aastu.edu.et 19-May-25
4 Probability Mass Function(Pmf)
 Let X be discrete RV and f be a function with 𝒇 𝒙 = 𝑷 𝑿 = 𝒙 for all x in
range of X(i.e R). If 𝒇 𝒙 ≥ 𝟎 𝒙 𝒊𝒏 R and σ𝒙 𝒇 𝒙 = 𝟏 , then 𝑓 𝑥 is called
A
probability mass function of X.
 Example Find the pmf of the number of heads observed in three tosses of
coin.
 In three tosses of a coin, X = Number of heads observed
✓ S ={HHH, HHT, HTH, HTT, THH, THT, TTH, TTT}
✓ X(HHH) =3,
✓ X(HHT)=X(HTH)=X(THH)=2 PMF of X
✓ X(TTH)=X(THT)=X(HTT)=1 𝑥 0 1 2 3
𝑓(𝑥) = 𝑃(𝑋 = 𝑥) 1 3 3 1
✓ X(TTT)=0
8 8 8 8
✓ X(S) = {0,1,2,3}
seid.belay@aastu.edu.et 19-May-25
5 Probability Mass …
Example: I have an unfair coin for which P(H)=p, where 0<p<1. I toss the coin
repeatedly until I observe a heads for the first time. Let Y be the total number
of coin tosses. Find the distribution of Y.
 Possible values of Y are 1, 2, 3, …
 P(Y=1) =P(H)? P
 P(Y=2) =P(TH) ? (1-p)p
… …
 P(Y=k) = P(TT…TH)=?
(1−p) (𝑘−1) p
 …check if it is PMF

seid.belay@aastu.edu.et 19-May-25
6 Probability Density Function(pdf)
𝐛
 Let X be continuous RV and 𝑓 be a function such that 𝐏 𝐚 < 𝐗 < 𝐛 = ‫𝐱𝐝 𝐱 𝐟 𝐚׬‬
𝑓𝑜𝑟 𝑎𝑙𝑙 𝑎 < 𝑏. The function f is called probability density function of the random
variable X IF IT SATISFIES .
i. 𝑓 𝑥 ≥ 0; 𝑎𝑛𝑑
∞
ii. ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥 = 1

𝑘. 𝑒 −3𝑥 𝑓𝑜𝑟𝑥 > 0

Example: If 𝑋 is the probability density 𝑓 𝑥 = ቊ
0 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
o what is the value of k? K=3

o Compute P(.5<x<1) 𝑒 −3/2 − 𝑒 −3

Example: Show that𝑓 𝑥 = 3𝑥 2 for 0 < 𝑥 < 1 represent a density function?

1
1. its Non Negative, 2. ‫׬‬0 3𝑥 2 𝑑𝑥 = 1 its Pdf

 Note: If X is a continuous RV and a & b are real constants with 𝑎 ≤ 𝑏, then

𝑃 (𝑎 ≤ 𝑥 ≤ 𝑏) = 𝑃 (𝑎 < 𝑥 ≤ 𝑏) = 𝑃 (𝑎 ≤ 𝑥 < 𝑏) = 𝑃 (𝑎 < 𝑥 < 𝑏)
seid.belay@aastu.edu.et 19-May-25
7 Example
1. Let the probability density function of a random variable be
𝑥 𝑓𝑜𝑟 0 ≤ 𝑥 < 1/2
1
𝑓 𝑥 = 𝑐 2𝑥 − 1 𝑓𝑜𝑟 ≤ 𝑥 < 1
2
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
 The value of c=_____________________
 The P(1/4<x<3/4)= ___________________

2
2. Let X be a continuous random variable with pdf 𝑓 𝑥 = ቊ3𝑥 𝑖𝑓 0 < 𝑥 < 1
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
2 1
Find 𝑃(𝑋 < |𝑋 > )
3 3

seid.belay@aastu.edu.et 19-May-25
8 Distribution Function
 Let X be a RV with pdf/or pmf/ f, then the function
𝐹(𝑥) = 𝑃(𝑋 ≤ 𝑥) is called Distribution Function/ or cumulative probability distribution function.
෍ 𝑓(𝑥𝑖) 𝑓𝑜𝑟 𝑝𝑚𝑓 𝑓
𝑥𝑖 ∈𝑋,𝑥𝑖 ≤𝑥
𝐹 𝑥 = 𝑃 𝑋 ≤𝑥 = 𝑥
න 𝑓 𝑥 𝑑𝑥 𝑓𝑜𝑟 𝑝𝑑𝑓 𝑓
−∞

Example: Find F(x) for

1. The number of heads in three tosses of a balanced coin 0 𝑓𝑜𝑟 𝑥𝜖(−∞, 0)
1
0 1 2 3 𝑓𝑜𝑟 𝑥𝜖[0,1)
8
𝑓(𝑥) = 𝑃(𝑋 = 𝑥) 1 3 3 1 ➔ 4
𝐹 𝑥 = 𝑓𝑜𝑟 𝑥𝜖[1,2)
8
8 8 8 8 7
𝑓𝑜𝑟 𝑥𝜖[2,3)
8
1 𝑓𝑜𝑟 𝑥 ∈ [3, ∞)
2. Random variable defined by 𝑓 𝑥 = 3𝑥 2 for 0 < 𝑥 < 1 0 elsewhere 0 𝑥<0
𝐹 𝑥 = ቐ𝑥 3 𝑥 ∈ [0,1)
seid.belay@aastu.edu.et
1 19-May-25
𝑥>1
9
Distribution Function …
 Properties of F(x)
 0 ≤ 𝐹(𝑥) ≤ 1 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑥 𝑖𝑛 𝑅
 𝐹(𝑥) is non-decreasing [i.e.,𝐹(𝑥) ≤ 𝐹(𝑦) 𝑖𝑓 𝑥 ≤ 𝑦].
 𝐹(𝑥) is continuous from the right [i.e., lim+ 𝐹(𝑥 + ℎ) = 𝐹(𝑥) for all 𝑥]
ℎ՜0

 lim 𝐹(𝑥) = 0 and lim 𝐹(𝑥) = 1

𝑥՜−∞ 𝑥՜∞

 P (a ≤ x ≤ b) = F(b) - F(a) for any real two constants a and b with a ≤ b, and
𝑑𝐹(𝑥)
 𝑓 𝑥 = where the derivative exist
𝑑𝑥

seid.belay@aastu.edu.et 19-May-25
10 Introduction to Expectation: Mean, Variance
 The expectation of a random variable X is often called mean and denoted by
E(X) which gives the long run average value of X.

σ𝑥∈𝑋 𝑥 𝑃(𝑋 = 𝑥) 𝑓𝑜𝑟 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒 𝑟𝑣 𝑋

 It is defined as 𝐄 𝐱 = ቐ ∞
‫׬‬−∞ 𝑥 𝑓 𝑥 𝑑𝑥 𝑓𝑜𝑟 𝑐𝑜𝑛𝑡𝑖𝑛𝑜𝑢𝑠 𝑟 𝑣 𝑋

 Example: If you are a financial analyst in a development company and your

manager just asked you to assess the viability of future development projects
and select the most promising one. According to estimates, up on completion
projects are valued as : Project A Project B
 Which project do you chose? Value Pr Value Pr
$2,000,000 0.4 $3,000,000 0.3
$500,000 0.6 $200,000 0.7

seid.belay@aastu.edu.et 19-May-25
11 Examples

 A game involves rolling a Korean die (4 faces). If a one, two, or three shows,
the player receives the face value of the die in dollars, but if a four shows,
the player is obligated to pay $4. What is the expected value of the game?

 During her four years at college, Niki received A's in 30% of her courses, B's
in 60% of her courses, and C's in the remaining 10%. If A = 4, B = 3, and C =
2, find her grade point average.

seid.belay@aastu.edu.et 19-May-25
12 Example
 The distribution of the amount of gravel (in tons) sold by a
particular construction supply company in a given week is a
3
1 − 𝑥 2 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1
continuous RV X with pdf 𝑓 𝑥 = ൝2
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

𝑓(𝑥)
 Find the expected amount of gravel sold.

1.5
1 3 3 1
Sol. 𝐄 𝐱 = ‫׬‬0 𝑥 × 1− 𝑥2 𝑑𝑥 = ‫׬‬ 𝑥 − 𝑥 3 𝑑𝑥
2 2 0
3 𝑥2
1 3 1 1𝑥4
= × = × − = 3/8
−
2 0 2
2 2 44
0 1 x
The expected amount of gravel sold in a given week is 3/8 tons.

seid.belay@aastu.edu.et 19-May-25
13 Example
Example: The time it takes to repair a personal computer is a random variable
whose density, in hours, is given by
1
𝑓 𝑥 = ቐ2 𝑖𝑓 0 < 𝑥 < 2
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
The cost of the repair depends on the time it takes and is equal to 40 + 30√𝑥
when the time is x. Compute the expected cost to repair a personal
computer.

seid.belay@aastu.edu.et 19-May-25
14 Properties of expectation
 For random variables X, and Y, and constants k, a, and b
 If X = K, E(X) = K
 E(aX + b) = a E(X) + b
 E(X ± Y) = E(X) ± E(Y).
 𝐄(𝐗𝐘) = 𝐄(𝐗)𝐄(𝐘) if X and Y are independent

 Example: If two random variables X and Y have density functions

1 1
𝑓 𝑥 = 8 𝑥 for 0 < 𝑥 < 4 and 𝑓 𝑦 = 12 𝑦 for 1 < y <5.
Find a. E (X) b. E(Y) C. E(X + Y) d. E(2X + 3Y)
e. if X and Y are independent, E(XY) =?

seid.belay@aastu.edu.et 19-May-25
 Variance
15  Let X is a random variable. The variance of X, denoted by 𝑉(𝑋) 𝑜𝑟 𝑉𝑎𝑟(𝑋) or 𝛿𝑥2 , defined
2.
as: 𝑉 𝑋 = 𝐸 𝑋 – 𝐸 𝑋 The positive square root of 𝑉(𝑋) is called the
𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑋 and denoted by 𝜎𝑥 .

2 σ𝑥∈𝑋 𝑥 2 𝑃(𝑋 = 𝑥) 𝑓𝑜𝑟 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒 𝑟𝑣 𝑋

 Var(X) = 𝐸 𝑋 – 𝐸 𝑋 =𝐸 𝑥 2
− 𝐸(𝑥) ; 𝐄 𝒙
2 𝟐
=ቐ ∞ 2
‫׬‬−∞ 𝑥 𝑓 𝑥 𝑑𝑥 𝑓𝑜𝑟 𝑐𝑜𝑛𝑡𝑖𝑛𝑜𝑢𝑠 𝑟 𝑣 𝑋
Example Find the variance of the amount of
Example: The number of packets of an item sold gravel sold in a given week.
in a day is a random variable as shown below 3
1 − 𝑥 2 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1
Packets(x) 3 4 5 𝑓 𝑥 = ൝2
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Probability 0.3 0.6 0.1 Sol. Var X = 𝐸 −𝐸 𝑥 𝑥2 2

Find Variance of X. We saw that 𝑬 𝒙 =3/8

1 3 3 1 2
Solution X 3 4 5 Sum
𝐸 𝑥 2 = ‫׬‬0 𝑥 2 2
1 − 𝑥 2 𝑑𝑥 = ‫׬‬
2 0
𝑥 − 𝑥 4 𝑑𝑥

3 𝑥 3 5
𝑥 1 3 1 1
P .3 .6 .1 1 = 2× − = × − =1/5
3 5 0 2 3 5
X*p .9 2.4 .5 3.8 1
Therefore, Var X = 𝐸 𝑥2 − 𝐸 𝑥 2 = −
𝐱 𝟐 ×P 2.7 9.6 2.5 14.6 5
3 2 19
E(X) =3.8 E(𝐱 𝟐 ) =14.6 Var(x) =0.16 = 320 ≈ 0.06
seid.belay@aastu.edu.et 8
Standard deviation = 𝑉𝑎𝑟(𝑋) = 0.06 ≈ 0.243
19-May-25
16 Properties of variance
 For a constant k, random variables X and Y with finite variance,
 Var(X + K) = Var(X)
 Var(KX) = K2 Var(X)
 𝑽𝒂𝒓 𝑿 ± 𝒀 = 𝑽𝒂𝒓 𝑿 + 𝑽𝒂𝒓(𝒀) When X and Y are independent
 For any real a , Var(X) = E[(X – a)2] – [E(X) – a]2.
 Example:
1. If X is RV Satisfying E[(X – 1)2] = 10, E[(X – 2)2] = 6.
Find (a) E(X) (b) Var(X) (c) Standard deviation of X?

3𝑥 2 𝑖𝑓 0 < 𝑥 < 1
2. Let X be a continuous random variable with pdf 𝑓 𝑥 = ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
2
If 𝑌 = 𝑋 + 3 find 𝑉𝑎𝑟 (𝑌)
seid.belay@aastu.edu.et 19-May-25
17 Moments and Moment Generating Functions(MGF)
 Let X be a random variable with pmf/ or pdf f: ෍ 𝒙𝒏 𝒇 𝒙 𝒊 𝒇𝒐𝒓 𝒑𝒎𝒇 𝒇
 The n-th moment about origin is X is defined as: 𝝁′ 𝒏 = 𝑬 𝑿𝒏 = 𝒊=𝟏
∞
න 𝒙𝒏 𝒇 𝒙 𝒅𝒙 𝒇𝒐𝒓 𝒑𝒅𝒇 𝒇
−∞
 Moment Generating Function of X is Denoted by 𝑴𝑿 (𝒕). Defined as: 𝑴𝑿 𝒕 = 𝑬 𝒆𝒕𝒙
𝒕𝟐 𝒙𝟐𝒊 𝒕𝒏 𝒙𝒏
σ𝒊=𝟏 𝒆𝒕𝒙𝒊 𝒇 𝒙𝒊 = σ𝒊=𝟏 𝟏 + 𝒕𝒙𝒊 + + ⋯+ 𝒊
+ ⋯ 𝒇 𝒙𝒊 , 𝒇 𝒊𝒔 𝒑𝒎𝒇
𝟐! 𝒏!
𝑬 𝒆𝒕𝒙 =
∞ ∞ 𝒕𝟐 𝒙𝟐 𝒕𝒏 𝒙𝒏
‫׬‬−∞ 𝒆𝒕𝒙 𝒇 𝒙 𝒅𝒙 = ‫׬‬−∞ 𝟏 + 𝒕𝒙 +
𝟐!
+ ⋯+
𝒏!
+ ⋯ 𝑓 𝑥 𝑑𝑥 , 𝒇 𝒊𝒔 𝒑𝒅𝒇
𝑡 2 𝝁′ 2 𝑡 𝑛 𝝁′ 𝑛
=1+ 𝑡𝝁′ 1 + 2! + ⋯ + 𝑛! + ⋯

𝒅𝒏 𝑴𝑿
✓ | = 𝝁′ 𝑛 = 𝑬 𝑿𝒏
𝒅𝒕𝒏 𝒕=𝟎

✓ 𝑴𝑿−𝝁 𝒕 = 𝑬 𝒆𝒕(𝒙−𝝁) = 𝒆−𝒕𝝁 𝑬 𝒆𝒕𝒙 = 𝒆−𝒕𝝁 𝑴𝑿 𝒕 , 𝒊. 𝒆. 𝜇𝑛 = 𝒆−𝒕𝝁 𝝁′ 𝑛

✓ 𝑴𝑿 𝒕 uniquely determines distributions. 𝐢. 𝐞. 𝑴𝑿 𝒕 = 𝑴𝒀 𝒕 ⟹ 𝑭𝑿 𝒙 = 𝑭𝒀 𝒙

seid.belay@aastu.edu.et 19-May-25
18 Common Probability
Distributions
 Binomial Distribution
 Poisson Distribution
 Poisson Approximation to Binomial
 Gaussian/Normal/ Distribution
 T-Distribution
 Exponential Distribution
 Chi-square Distribution
 F-Distribution

seid.belay@aastu.edu.et 19-May-25
19 Bernoulli & Binomial Distributions
 Bernoulli experiment:
 Consists of a random experiment with two possible outcomes; (a success or a failure)
 The probability of a success, denoted by p, probability of failure=1-p .
 X=The number of successes
o 𝑷 𝑿 = 𝒙 = 𝒑𝒙 𝟏 − 𝒑 𝟏−𝒙 ; 𝑥 = 0, 1
 X 1 0 o E(x) = P, Var(x) = p(1-p)
P(X=x) p 1-p o 𝑴𝑿 𝒕 = 𝟏 − 𝒑 + 𝒑𝒆𝒕
 Called Bernoulli Distribution Mean=p, Variance= p(1-p)
 Binomial Distribution
 Random experiment with two outcomes - a success with probability p, or a failure with 1-p prob.
 Consists of n repeated independent trials; p remains constant from trial to trial.
 X=The number of successes in n trials, called binomial variable & take values 0,1, …, n
𝒏 𝒙 𝒏−𝒙
Binomial Distribution o 𝑷 𝑿=𝒙 =
𝒙
𝒑 𝟏−𝒑 ; 𝑥 = 0, 1, … , 𝑛

o E(x) = 𝒏𝑷, Var(x) = np(1-p)

seid.belay@aastu.edu.et 19-May-25
o 𝑴𝑿 𝒕 = 𝟏 − 𝒑 + 𝒑𝒆𝒕 𝒏
20 Binomial Distribution:- Examples
❖ Example1: Hospital records show that of patients suffering from a certain disease,
75% die of it. What is the probability that of 6 randomly selected patients, 4 will
recover?
o Sol. Let X = Number of recovering patients, n = 6 random(independent), Success = recovery,
Probability of recovery = .25 this is binomial.
6
o Required: 𝑃 𝑋 = 4 = .254 1 − .25 6−4
=15×2.1973×10−3=0.0329595
4

❖ Example 2: Nine percent of men and 0.25% of women cannot distinguish between the
colors red and green. This is the type of color blindness that causes problems with traffic
signals. If six men are randomly selected for a study of traffic signal perceptions, find the
probability that exactly two of them cannot distinguish between red and green.

seid.belay@aastu.edu.et 19-May-25
21 Binomial …
❖ Example 3: The probability that a student passes statistics exam is 0.8. If there are
50 students in a class, what is the probability that

a. Exactly 50% of the students pass the exam

b. At least 60% of the students pass the exam
c. At most 80% of students pass the exam

seid.belay@aastu.edu.et 19-May-25
22 Poisson Distribution
 Developed by the French mathematician Simeon Denis Poisson in 1837.
 A Poisson variable Counts number of events occurring in time interval/space and satisfies:
 The number of successes in two disjoint time intervals is independent
 The probability of a success during a small time interval is proportional to the entire length of
the time interval
- number of typing errors on a page
 Applications:
- birth defects and genetic mutations
- number of car accidents …
 Probability Distribution of Poison RV:
𝒆−𝝀 𝝀𝒊
𝑷 𝑿=𝒊 = ; 𝒊 = 𝟎, 𝟏, 𝟐, 𝟑, …
𝒊!
𝒕
 Mean = E(x)= 𝝀 Variance =𝑬 𝒙𝟐 − 𝑬 𝒙 𝟐 =𝝀 𝑴𝑿 𝒕 = 𝒆𝝀(𝒆 −𝟏)
 Example: A life insurance salesman sells on the average 3 life insurance policies per week. Use
Poisson's law to calculate the probability that in a given week he will sell
a) Some policies b) at least 2 at most 5 policies c) Assuming that there are 5 working days per
week,seid.belay@aastu.edu.et
what is the probability that in a given day he will sell one policy? 19-May-25
23 Poisson …
 For a Poisson Variable X:

 Sum of Poisson Variables is also Poisson

𝒆−𝝀𝟏 𝝀𝒊𝟏
If 𝑷 𝑿=𝒊 = ;𝒊
= 𝟎, 𝟏, 𝟐, 𝟑, … and
𝒊!
𝒆−𝝀𝟐 𝝀𝒊𝟐
𝑷 𝒀=𝒋 = ; 𝒋 = 𝟎, 𝟏, 𝟐, 𝟑, …
𝒊!
Then, 𝒁 = 𝑿 + 𝒀
𝒆−𝝀 𝝀𝒌
𝑷 𝒁=𝒌 = 𝐰𝐡𝐞𝐫𝐞 𝝀 = 𝝀𝟏 + 𝝀𝟐 ; 𝐤 = 𝟎, 𝟏, 𝟐, …
𝒌!

 Your friend sends you text messages at a daily Poisson rate of 𝝀𝟏 =3 and your dad
independently sends you text massages at a daily rate of 𝝀𝟐 =1. suppose that you
don’t receive messages from other source. What is the probability that you will
receive exactly 5 messages tomorrow?
seid.belay@aastu.edu.et 19-May-25
24 Exercise
Let’s say we have two independent random Poisson variables for requests received at a
web server in a minute: X = # requests from humans/minute, X ∼ Poi(5) and Y = # requests
from bots/minute, Y ∼ Poi(3).

a) What is the probability of having 10 requests in a given minute?

b) Given that we have 10 requests in a given minute, what is the probability that 7 of the
requests are human requests?

seid.belay@aastu.edu.et 19-May-25
25 Poisson Approximation To Binomial
 For large n, calculating probabilities for binomial distribution is Nasty.
The Good Thing:
 In a binomial distribution if 𝑛 ՜ ∞, 𝑎𝑛𝑑 𝑝 ՜ 0, 𝑛𝑝 ՜ 𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝜆

𝑥 𝑛 −𝑥
𝑛 𝑥 𝑛−𝑥
𝑛 × 𝑛 − 1 … (𝑛 − 𝑥 + 1) 𝜆 𝜆 𝜆
lim 𝑝 1−𝑝 = lim 1− 1−
𝑛՜∞ 𝑥 𝑛՜∞ 𝑥! 𝑛 𝑛 𝑛
x=terms
𝐵𝑖𝑛𝑜. 𝐷𝑖𝑠𝑡.
𝜆𝑥 𝑛× 𝑛−1 …(𝑛−𝑥+1) 𝜆 𝑛 𝜆 −𝑥
= × lim × lim 1− × lim 1 − =
𝑥! 𝑛՜∞ 𝑛𝑥 𝑛՜∞ 𝑛 𝑛՜∞ 𝑛

𝜆𝑥 𝑒 −𝜆
≅1 ≅ 𝑒 −𝜆 ≅1
≅
𝑥!
x=terms
𝑃𝑜𝑖𝑠. 𝐷𝑖𝑠𝑡.

Note: The approximation is Good when

n > 20, np < 5 OR n(1-p) < 5
19-May-25
seid.belay@aastu.edu.et
26 Poisson Approximation …
Example1: Suppose that 1 in 5000 light bulbs are defective. What is the probability that
one can get at most 3 defective bulbs from a shipment of 10000 bulbs.
 Sol. Let X = number of defective bulbs. P= probability of defective =
0.0002. this is binomial variable. But n is large and p is small, hence we
can approximate it by Poisson. 𝜆 = np =2
P(X<=3) = P(X=0)+P(X=1)+P(X=2)+P(X=3)
20 𝑒 −2 21 𝑒 −2 22 𝑒 −2 23 𝑒 −2
= + + + = 0.857123
0! 1! 2! 3!

Example2: A 747 plane has 360 seats. An airline has found that 1% of people with
tickets do not turn up. Find the probability that more than 4 people do not turn up for a
flight.

seid.belay@aastu.edu.et 19-May-25
27 Gaussian Distribution / Normal Distribution
 The most important probability distribution in statistics – It fits many phenomenon
 Important in hypothesis testing, regression, inference
 It is symmetric, bell shaped, asymptotic to x-axis.
 Mean = Mode = Median
 For normal random variable X with mean 𝜇 and variance 𝜎 2
𝒙−𝝁 𝟐 𝟏
𝟏 − 𝝁𝒕+𝟐𝒕𝟐 𝝈𝟐
PDF: 𝒇 𝒙 = 𝒆 𝟐𝝈𝟐 , 𝑥 ∈ (−∞, ∞) With MGF 𝑴𝑿 𝒕 = 𝒆
𝝈 𝟐𝝅

seid.belay@aastu.edu.et 19-May-25
28 Standard Normal Distribution
 A special normal distribution where the mean is 0 and the
standard deviation is 1.
𝑧 2
1 −2
 Pdf: 𝑓 𝑧 = 𝑒 , 𝑧 ∈ (−∞, ∞)
2𝜋
𝑋− 𝜇
 Every Normal distribution can be transformed to it with 𝑍 = 𝜎

 Finding Areas Under the Curve of a Normal Distribution

The proportion of the area that falls under the curve between
two points on a probability distribution plot indicates the
probability that a value will fall within that interval
 Example: The diameter of certain variety of tree follows
normal distribution with mean15cm and a standard
deviation of 3cm. a) Find The proportion of trees with
12 − 15
diameter less than 12. Z = 3 = −1
proportion = P(d<12) = P(z<-1) = P(z>1) =1-p(z<1) =0.1587
seid.belay@aastu.edu.et 19-May-25
29

seid.belay@aastu.edu.et 19-May-25
30 Example (Cont’d)
 Find the probability that diameter (d)
i. Exceeds 17.5cm
ii. Falls between 11cmand 14cm Z
iii. Falls between 12cm and 18cm
iv. The third quartile value q
Sol.
17.5−15
o 𝑃 𝑑 > 17.5 = 𝑃 𝑧 >
3
= 𝑃(𝑍 > 0.83) Z
11−15 14−15
o 𝑃 11 < 𝑑 < 14 = 𝑃 <𝑧<
3 3

= 𝑃(−1.33 < 𝑧 < −0.33)

12−15 18−15
o 𝑃 12 < 𝑑 < 18 = 𝑃 <𝑧<
3 3
Z
= 𝑃 −1 < 𝑧 < 1 = 2 ∗ 𝑃(0 < 𝑧 < 1)
o 3rd quartile means the value below which 75% of
the observations fall. P(d<q3)= 0.75

seid.belay@aastu.edu.et 19-May-25 d
31 Example
 The score of students in an exam is normally distributed with mean 75 and standard deviation of
10. If the professor decided to assign grade A for the top 10% scorers, what is the minimum score to
get Grade A?

 The probability that a river flow exceeds 2,000 cubic meters per second is 12.3%. The coefficient of
variation of these flows is 20%. Assuming a normal distribution, calculate
a) the mean of the flow.
b) the standard deviation of the flow.
c) the probability that the flow will be between 1300 and 1900 m3/s

 The Scholastic Aptitude Test mathematics test scores across the population of high school seniors
follow a normal distribution with mean 500 and standard deviation 100. If five seniors are randomly
chosen, find the probability that
a) all scored below 600 and
b) exactly three of them scored above 640

seid.belay@aastu.edu.et 19-May-25
32 Normal Approximation to Binomial
 Consider Tossing of a coin balanced 20 times
and consider the number of heads observed(x)
 Considering the binomial,
20
P(X=10) = 0.510 1 − 0.5 20−10 = 0.176197
10
The shape of the distribution very close to normal
shape

Using Normal To approximate Binomial

 𝜇 = 𝑛𝑝 = 10 𝑎𝑛𝑑 𝜎 2 = 𝑛𝑝 1 − 𝑝 = 5
 Continuity correction factor 0.5
9.5−10 10.5−10
 𝑃 9.5 < 𝑋 < 10.5 = 𝑃 5
<𝑧< 5
=
𝑃 −0.2236 < 𝑧 < 0.2236 = 0.176937
Normal Approximation is good when both np and
np(1-p) are greater than 5
seid.belay@aastu.edu.et 19-May-25
33 Important Result/Sum of INDV is Normal/
 Let 𝑋1 , 𝑋2 , …, 𝑋𝑛 be independently and normally distributed random variables with
means 𝜇𝑖 and variance 𝜎𝑖2 for 𝑖 = 1,2, … , 𝑛, then Z = σ𝑛𝑖=1 𝑋𝑖 is also normal variable
Proof
σ𝑛
𝑀𝑍 𝑡 = 𝐸 =𝐸𝑒 𝑡𝑧 𝑒 𝑖=1 𝑿𝒊
𝑡
❖ Z has the same moment generating
= 𝐸 𝑒 𝑡𝑿𝟏 × ⋯ × 𝑒 𝑡𝑿𝒏 function as a normal random variable
having mean 𝜇 = σ𝑛𝑖=1 𝝁𝒊 and variance
= 𝐸 𝑒 𝑡𝑿𝟏 × ⋯ × 𝐸 𝑒 𝑡𝑿𝒏 𝜎 2 = σ𝒏𝒊=𝟏 𝝈𝟐𝒊
𝟏 𝟏
𝑡𝝁𝟏 +𝟐𝒕𝟐 𝝈𝟐𝟏 𝑡𝝁𝒏 +𝟐𝒕𝟐 𝝈𝟐𝒏
=𝑒 × ⋯× 𝑒
𝟏 𝟐 𝒏
𝑡 σ𝑛 𝝁 𝒊 + 𝒕 σ 𝝈 𝟐 ❖ By one-to-one correspondence
=𝑒 𝑖=1 𝟐 𝒊=𝟏 𝒊
between moment generating
functions and distributions, we can
MGF for Normal Variable with: conclude that σ𝑛𝑖=1 𝑋𝑖 is normal with
Mean: 𝜇 = σ𝑛𝑖=1 𝝁𝒊 mean 𝝁 and variance 𝝈𝟐
Variance: 𝜎 2 = σ𝒏𝒊=𝟏 𝝈𝟐𝒊
=> Z is normally distributed
seid.belay@aastu.edu.et 19-May-25
34 Example
 A report indicate that the yearly precipitation in Los Angeles is a normal random variable
with a mean of 12.08 inches and a standard deviation of 3.1 inches.
a) Find the probability that the total precipitation during the next 2 years will exceed 25
inches.

b) Find the probability that next year’s precipitation will exceed that of the following
year by more than 3 inches. Assume that the precipitation totals for the next 2 years
are independent.

seid.belay@aastu.edu.et 19-May-25
35 Example
 History suggests that scores on the Math portion of the Standard Achievement Test
(SAT) are normally distributed with a mean of 529 and a variance of 5732. History
also suggests that scores on the Verbal portion of the SAT are normally distributed
with a mean of 474 and a variance of 6368. Select two students at random.
Let X denote the first student's Math score, and let Y denote the second student's
Verbal score. What is P(X>Y)?

0.6915

seid.belay@aastu.edu.et 19-May-25
36 Exponential Distribution
 A random variable X is said to be
exponentially distributed with rate 𝜆 if its pdf is
−𝝀𝒙
𝒇 𝒙; 𝝀 = ቊ 𝝀𝒆 𝒙≥𝟎
𝟎 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆
 Its Mean and Variance respectively:
1 1
𝑎𝑛𝑑
𝜆 𝜆2

 Used to model time to an event to occur.

𝝀
 MGF 𝑴𝑿 𝒕 = 𝝀−𝒕 , 𝝀 > 𝒕

Example: The time (in hours) required to repair a machine is an exponentially distributed
random variable with parameter λ = 1.
(a) What is the probability that a repair time exceeds 2 hours?
(b) What is the conditional probability that a repair takes at least 3 hours, given that its
duration exceeds 2 hours?
seid.belay@aastu.edu.et 19-May-25
37 T-Distribution
 It’s a distribution very similar to standard
normal distribution
− 𝝊+𝟏
𝒙𝟐 𝟐
𝟏+
 Its pdf:
𝝊
𝑓 𝑥 = ; 𝑥 𝑖𝑠 𝑟𝑒𝑎𝑙
𝐵 0.5,0.5𝜐 𝝊

𝟏
 Where 𝑩 𝜶, 𝜷 = ‫𝜶𝒕 𝟎׬‬−𝟏 𝟏 − 𝒕 𝜷−𝟏 𝒅𝒕 is beta
function
 Mean= Mode =Median = 0
𝜈
 Variance = ;𝜈 >2
𝜈−2
Example: If T has a t-distribution with 8 degrees
of freedom, find
(a) P{T ≥ 1.108}
(b) P{T ≤ 1.860}, and
(c) P{-1.108 < T < 1.108}
seid.belay@aastu.edu.et 19-May-25

𝑃 𝑇15=𝑑𝑓 < 1.074 = 0.85

 38 Chi-Squared Distribution(𝝌𝟐 )
 Let 𝑥1 , … , 𝑥𝑛 be random sample with 𝑥𝑖 ~𝑁(𝜇𝑖 , 𝜎𝑖2 ) then
𝒙𝒊 −𝝁𝒊 𝟐
the statistic 𝝌𝟐 = 𝒏
σ𝒊=𝟏 is called Chi-squared
𝝈𝒊
𝒙𝒊 − 𝝁 𝟐 σ𝒏
𝒊=𝟏 𝒙𝒊 −𝝁
𝟐
Variable. If IID Normal, 𝝌𝟐 = 𝒏
σ𝒊=𝟏 =
𝝈 𝝈𝟐
−𝑥 𝜈
−1
 Its pdf: 𝑓 𝑥 =
𝑒 2 𝑥2
𝜈
𝜈
𝑓𝑜𝑟 𝑥 ≥ 0 ( n − 1) s 2
=
 ( x − x) 2

= 2
22 Γ
2
 2
 2

∞
 Where Γ 𝑎 = ‫׬‬0 𝑡 𝑎−1 𝑒 −𝑡 𝑑𝑡 𝑖𝑠 𝑔𝑎𝑚𝑚𝑎 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
 Mean = 𝜈 = Degree of Freedom = n-1
 Variance = 2 𝜈
 𝑃 𝝌𝟐 > 𝝌𝟐𝟎.𝟏,𝟗 = 𝟎. 𝟏, 𝒘𝒉𝒆𝒓𝒆 𝝌𝟐𝟎.𝟏,𝟗 = 𝟏𝟒. 𝟔𝟖𝟒
 𝑃 𝝌𝟐 > 𝝌𝟐𝟎.𝟗𝟕𝟓,𝟔 = 𝟎. 𝟗𝟕𝟓 where 𝝌𝟐𝟎.𝟗𝟕𝟓,𝟔 = 𝟏. 𝟐𝟑𝟕
 Use Case: seid.belay@aastu.edu.et 19-May-25

 Test of independence, Goodness of fit

 F-table of critical Values of 𝜶 = 𝟎. 𝟏, 𝑭𝒐𝒓 𝑭(𝒅𝒇𝟏, 𝒅𝒇𝟐)
39 F- Distribution
 Consider any two normally distributed populations 𝐴 and 𝐵.
 Let
𝑆𝐴2 = variance of a sample of size 𝑛𝐴 drawn from 𝐴 and
𝑆𝐵2 = variance of a sample of size 𝑛𝐵 drawn from 𝐵
𝝌𝟐
𝑺𝟐𝑨 ൗ𝒏 −𝟏
 𝑇ℎ𝑒𝑛, 𝑭 = = 𝝌𝟐 𝑨
called F-statistic which follows F-
𝑺𝟐𝑩 ൗ𝒏 −𝟏
𝑩
Distribution with 𝑛𝐴 -1 and 𝑛𝐵 -1 degree of Freedom.
 For each upper-tail probability 𝛼 , there is a critical value F
such that  = Pr F  F   

seid.belay@aastu.edu.et 19-May-25
49

Thank You All!

seid.belay@aastu.edu.et 19-May-25