TM1151 - Business Statistics and Software Applications

Chapter 05 – Distribution Theory

1
Learning Outcomes

At the end of this lesson you will be able to,

1. Apply basic statistics in analyzing business problems.

2
Random Variables
• A random variable (r.v.) X is a function defined on a sample space (S),
that associates a real number, X(ω) = x, with each outcome ω in S.

• Simple Definition : A random variable is a mapping between all the

outcomes in the sample space with a set of real numbers.

• Random variables are denoted by using capital letters

3
 Random
Variables

Discrete Continuous
Random Random
Variables Variables

4
Discrete Random Variables
• A random variable is said to be discrete, if it can assume only distinct values.

OR
• In other words, it can assume only countable number of values.

Examples
1. Toss a coin 5 times. Let X be the number of heads appeared.
Then, X - 0,1,2,3,4,5
1. Roll a die twice. Let X be the number of times 4 comes up. X - 0,1,2
2. Suppose we toss two coins. Assume that all the outcomes are equally likely
(fair coins). Let Y be the number of heads appeared. Then, Y - 0,1,2
5
PROBABILITY DISTRIBUTIONS

• The set of all ordered pairs (𝑥, Pr(𝑋 = 𝑥)) of a discrete r.v. (𝑋), is known
as the probability distribution

• This is also known as the probability mass function (p.m.f.) and is denoted
by 𝑃𝑋(𝑥).

• Simple Definition: All the possible values of a discrete random variable

with their corresponding probability values is known as probability
distribution.
6
Properties
• 𝑃𝑋(𝑥) refers to 𝑃(𝑋 = 𝑥).

• The probability distribution function is always non-negative.

• P
all x
X
( x)  1

• The cumulative distribution function (c.d.f.) of the random variable 𝑋

is defined by 𝐹𝑋 (𝑥) = Pr(𝑋 ≤ 𝑥)

7
Example 01
Suppose we toss two coins. Assume that all the outcomes are equally
likely (fair coins). Let Y be the number of heads appeared. Calculate
the cdf,

Y 0 1 2
Pr(Y=y) 1/4 1/2 1/4

8
Expected Value & Variance

• This is same as mean of the random variable.

• Let X be a discrete random variable with p.m.f. 𝑃𝑋(x).

• Then the expected value of X, denoted by E(X), is defined by

𝑬 𝑿 = 𝒂𝒍𝒍 𝒙 𝒙 ∗ 𝐏𝐫(𝑿 = 𝒙)

• The variance of a random variable X is defined by

𝑽 𝑿 = 𝑬(𝑿 − 𝑬(𝑿))𝟐 = 𝑬 𝑿𝟐 − [𝑬 𝑿 ]𝟐
9
Properties [E(X)]
• Let X & Y be two random variables.

• Then,
• 𝐸(𝑐) = 𝑐

• 𝐸 𝑔(𝑋) = 𝑎𝑙𝑙 𝑥 𝑔(𝑋) ∗ Pr(𝑋 = 𝑥)

• 𝐸[𝑔(𝑋) + 𝑐] = 𝐸[𝑔(𝑋)] + 𝑐

• 𝐸[𝑐 ∗ 𝑔(𝑋)] = 𝑐 ∗ 𝐸[𝑔(𝑋)]

• 𝐸[𝑋 + 𝑌] = 𝐸[𝑋] + 𝐸[𝑌]

10
Properties [V(X)]
• Let X & Y be two random variables.
• Then,
• 𝑉(𝑐) = 0

• 𝑉[𝑔(𝑋) + 𝑐] = 𝑉[𝑔(𝑋)]

• 𝑉 𝑐∗𝑔 𝑋 = 𝑐 2 ∗ 𝑉[𝑔(𝑋)]

• 𝑉[𝑋 + 𝑌] = 𝑉[𝑋] + 𝑉[𝑌] + 2𝐶𝑜𝑣(𝑋, 𝑌)

• 𝑉[𝑋 − 𝑌] = 𝑉[𝑋] + 𝑉[𝑌] − 2𝐶𝑜𝑣(𝑋, 𝑌)

• If 𝑋 & 𝑌 are independent then, 𝐶𝑜𝑣(𝑋, 𝑌) = 0

11
Covariance
• Covariance is a measure of how the changes in one variable are
associated with the changes in second variable.

𝑪𝒐𝒗 𝐗, 𝐘 = 𝑿𝒊 − 𝑬 𝑿 𝒀𝒊 − 𝑬 𝒀 𝑷(𝑿𝒊 , 𝒀𝒊 )
𝒊=𝟏

12
Example 02
Suppose we toss two coins. Assume that all the outcomes are equally
likely (fair coins). Let Y be the number of heads appeared. Find E(Y)
and Var(Y).

Y 0 1 2
Pr(Y=y) 1/4 1/2 1/4

13
Example 03
I. To find out the prevalence of vaccine use, a researcher inquired
into the number of times a randomly selected 200 people aged 16
and over in an village had been vaccinated. He obtained the
following figures: never, 16 people; once, 30; twice, 58; three
times, 51; four times, 38; five times, 7. Assuming these proportions
continue to hold exhaustively for the population of that village,
what is the expected number of times those people in the village
had been vaccinated, and what is the standard deviation?

II. Let X and Y be two independent random variables. Suppose that we

know Var(2X−Y)=6 and Var(X+2Y)=9. Find Var(X) and Var(Y).

14
Discrete Probability Distributions

Discrete
Probability
Distributions

Bernoulli Binomial Poisson

Distribution Distribution Distribution

15
Bernoulli Distribution
• Only two possible outcomes (Success & Failure)
• Only one trial
Eg:- Tossing a coin once

𝑋 – Getting the success

𝑋 ~ 𝐵𝑒𝑟𝑛𝑜𝑢𝑙𝑙𝑖(𝑝)
𝑷𝑿 (𝒙) = 𝒑𝒙 (𝟏 − 𝒑) 𝟏−𝒙 [p.m.f.]
𝐸(𝑋) = 𝑝
𝑉(𝑋) = 𝑝(1 − 𝑝) 16
Binomial Distribution
• For each trial, only two possible outcomes (Success & Failure).

• No of trials (n) are fixed.

• Probability of success (p) is constant for each & every trial.

• Trials are independent.

• An expansion of the Bernoulli distribution.

• Each trial has a Bernoulli distribution.

Eg:- Tossing a coin 10 (n) times

17
𝑋 - The number of successes in n number of trials.
𝑋 ~ 𝐵𝑖𝑛(𝑛, 𝑝)

𝒏
𝑷𝑿 𝒙 = 𝒑𝒙 (𝟏 − 𝒑)𝒏−𝒙 ; 𝒙 = 𝟎, 𝟏, 𝟐, … , 𝒏 [𝒑. 𝒎. 𝒇. ]
𝒙

𝐸(𝑋) = 𝑛𝑝
𝑉(𝑋) = 𝑛𝑝(1 − 𝑝)

18
Poisson Distribution

• For each trial, only two possible outcomes (Success & Failure).

• Trials (n) is large.

• The occurrences are independent of each other [Assume that the

numbers of occurrences in disjoint Intervals] .

Eg:- Number of defects in a lot

19
𝑋 - The number of occurrences for a given rate of occurrence (λ)
𝑋 ~ 𝑃𝑜𝑖𝑠𝑠𝑜𝑛(𝜆)

𝒆−𝝀 𝝀𝒙
𝑷𝑿 (𝒙) = ; 𝒙 = 𝟎, 𝟏, 𝟐, … [𝒑. 𝒎. 𝒇. ]
𝒙!

𝐸(𝑋) = 𝜆
𝑉(𝑋) = 𝜆

20
Example 04
It is known that paper clips produced by a certain machine will be
defective with probability 0.01 independently of each other. If we
randomly pick 10 paper clips produced by this machine, what is the
probability that

a)exactly six paper clips will be defective?

b)at most 3 paper clips will be defective?
c)at least 2 paper clips will be defective?
d)What is the expected number of defectives?
e)What is the variance of defectives?

21
Example 05
Fifty beans seeds were planted and it is known that the probability of
any beans seed germinating is 0.4. Assuming that the beans seeds
have no other factors in germinating, find the following probabilities.
a)More than 10 beans seeds germinate.
b)More than 20 but fewer than 30 beans seeds germinate

Example 06
Suppose that, on average, in every two pages of a magazine there is
one typographical error. What is the probability of at least one error
on a certain page of the magazine?

22
Poisson Approximation

• If 𝑋 ∼ 𝐵𝑖𝑛(𝑛, 𝑝), then 𝑋 can be approximated with a Poisson

distribution with the rate parameter (𝜆) being equal to 𝑛𝑝 if 𝑝 is quite
small and 𝑛 is large.

• Usually this approximation can be used if 𝑝 < 0.1 and 𝑛 > 50.

23
Example 07
On a particular production line, the probability that an item is
defective is 0.001, determine the probability that out of 2000
products,

(a) exactly three and

(b) more than two products will be defective.

24
Continuous Random Variables

• A random variable is said to be continuous, if it can take any value

within a range.

• Continuous data are frequently measured in some way rather than

counted.

• If X is a continuous random variable, Pr(X = a) = 0 for any value of a.

Example: Temperature, Heart beat of a patient, Rainfall, Waiting time

for a bus
25
Probability Distributions

• For continuous random variables, the probability distribution cannot

be presented in a tabular form.

• Probability distribution function of a continuous random variable is

known as probability density function (pdf).

• The area under the p.d.f. gives probability values.

26
Definition

• The function 𝑓𝑋(𝑥) is a probability density function for the continuous

random variable 𝑋, defined over the set of real numbers (ℝ), if
• 𝑓𝑋(𝑥) ≥ 0,for all x ∈ R
∞
• 𝑓 (𝑥)
−∞ 𝑋
dx = 1

𝑏
• Pr 𝑎 < 𝑋 < 𝑏 = 𝑓 (𝑥)
𝑎 𝑋
dx

27
Properties
• Let X be a continuous random variable with a p.d.f. (𝑓𝑋(𝑥)), defined
over the set of real numbers (ℝ).
𝑥
• The c.d.f. 𝐹𝑋 (𝑥) = 𝑃𝑟(𝑋 ≤ 𝑥) = 𝑓 (𝑥)
−∞ 𝑋
dx
∞
• 𝐸[𝑔(𝑥)] = −∞
g(x) 𝑓𝑋(𝑥) 𝑑𝑥

• 𝑉[𝑔(𝑥)] = 𝐸[𝑔(𝑥) 2 ] – {𝐸[𝑔(𝑥)]} 2

28
Example 08
Suppose that the error in the reaction temperature, in °C, for a
controlled laboratory experiment is a continuous random variable 𝑋
having the probability density function,
𝑐𝑥 2 ; 0 ≤ 𝑥 ≤ 10
𝑓𝑋 𝑥 =
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
I. Find the value of c
II. Calculate Pr(0 < X ≤ 1)
III. Find the expected value and the variance
IV. Find the c.d.f.
29
 Continuous
Probability
Distributions

Exponential Normal
Distribution Distribution

30
Exponential Distribution
• The distribution is usually used to model life times.

Examples: duration between two customers at Bank ATMs, To model

patients entering to an accident ward.

X ~ Exp(𝝀)
𝜆𝑒 −𝜆𝑥 ; 𝑥 ≥ 0
𝑓𝑋 𝑥 =
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

𝐸(𝑋) = 1/ 𝜆
𝑉(𝑋) = 1/ 𝜆2
31
Example 09
The time, in hours, during which an electrical generator is operational is
a random variable that follows an exponential distribution with a mean
of 160. What is the probability that a generator of this type will be
operational for,

a) Less than 40 hours?

b) Between 60 and 160 hours?
c) More than 200 hours?

32
Normal Distribution
• This is most commonly used distribution. This is bell shaped
distribution and perfectly symmetric around µ.

𝑋 ~ 𝑁(µ, 𝜎 2 )

𝐸(𝑋) = µ
𝑉(𝑋) = 𝜎 2
33
• Normal distribution with µ = 0 and 𝜎 2 = 1 is known as the Standard
Normal Distribution.

• Evaluating probabilities with Normal requires complex integration.

• Statistical tables are used to simplify the procedure.

• Tables are only for the standard normal distribution.

Normal -> Standard Normal

𝑿−𝝁
If 𝑿 ~ 𝑵 ( µ, 𝝈𝟐 ), Then 𝒁 = ~ 𝑵 (𝟎, 𝟏)
𝝈

34
Example 10
I. For Z ∼ N(0, 1), calculate Pr(Z ≥ 1.13).

II. For X ∼ N(5, 4), calculate Pr(−2.5 < X < 1.13)

III. The actual marks for faculty of business of BPM students revealed
that they were normally distributed with a mean mark of 45 and a
standard deviation of 22. What is the probability that a randomly
chosen student will pass? (Assume that pass mark is 45)

35
Normal Approximation to Binomial Probabilities
• For 𝑋 ∼ 𝐵𝑖𝑛(𝑛, 𝑝) this approximation can be used if 𝑛 is large and 𝑝 is
moderate.
• A general rule can be defined as, np and n(1 − p) is greater than 5.
• Can be approximated with a r.v. with a distribution 𝑁(𝑛𝑝, 𝑛𝑝(1 − 𝑝)).
• A continuity correction is needed because a discrete distribution is
approximated with a continuous distribution.
• Continuity correction?

36
Example 11
Suppose that a sample of n = 1,600 tires of the same type are obtained
at random from an ongoing production process in which 8% of all such
tires produced are defective. What is the probability that in such a
sample 150 or fewer tires will be defective?

37
Normal Approximation to Poisson Probabilities

• If 𝑋 ∼ 𝑃𝑜𝑖𝑠𝑠𝑜𝑛(𝜆) then if 𝜆 is greater than 20, the approximation can

be used.

• Can be approximated with a r.v. with a distribution 𝑁(𝜆, 𝜆).

• A continuity correction is needed because a discrete distribution is

approximated with a continuous distribution (just as in the case of the
Binomial to Normal approximation)

38
Example 12
The annual number of earthquakes registering at least 2.5 on the
Richter Scale and having an epicenter within 40 miles of down town
Memphis follows a Poisson distribution with mean 22.5. What is the
probability that at least 25 such earthquakes will strike next year?

39
In Next Chapter…
• Sampling and sampling distributions will be discussed

40
Thank You

Rajika Gunarathne

Department of Management of Technology

Faculty of Business
University of Moratuwa

Email: rajikag@uom.lk
41