AST10113 Foundation Statistics

Lecture 5 & 6
Discrete Probability Distribution

Kirk Chan & Charmaine Lau

Foundation Statistics – Discrete Probability Distribution

 Learning Objectives
In this chapter, you learn:
 To compute probabilities from binomial,
hypergeometric, and Poisson distributions
 To use the binomial, hypergeometric, and Poisson
distributions to solve real life problems

Foundation Statistics – Discrete Probability Distribution 2

 Discrete Probability Distributions
 Probability Distribution can be classified as discrete and
continuous depends on the nature of the variable
 If a sample space contains a finite number, it is called a
Discrete Sample Space
 A random variable is called a Discrete Random Variable if
its set of possible outcomes is countable and finite
 Example:
A student receives a grade in a public examination
Outcomes: Discrete grades A, B, C, D, E, F or U
Sample space: Discrete (finite)
Random variable: Define
{X = 1} if rating A {X = 5} if rating E
{X = 2} if rating B {X = 6} if rating F
{X = 3} if rating C {X = 7} if rating U
{X = 4} if rating D
Foundation Statistics – Discrete Probability Distribution 3
 Continuous Random Variables
 If a sample space contains an infinite number of
possibilities, it is called a Continuous Sample Space
 A random variable is called a Continuous Random Variable
if it takes on values on a continuous scale
 Example:
Invest $10,000 in a blue chip
stock in Hong Kong.
Outcomes: Value of yield
Sample space: Continuous (infinite)
Random variable: Define X = value of yield
where 0  x  

Foundation Statistics – Discrete Probability Distribution 4

 Discrete vs Continuous Distributions
 Discrete Distribution  Continuous Distribution
– The distribution of a – Describes events over a
random variable X is continuous range, where
discrete in the sense that all the probability of a specific
its points are isolated points. outcome is zero.
– Its cumulative distribution – Its cumulative distribution
function only increases in function is continuous
jumps

Foundation Statistics – Discrete Probability Distribution 5

 Commonly known Probability Distributions
Probability Distributions

Discrete Continuous
Probability Distributions Probability Distributions

Bernoulli Uniform

Binomial Normal

Negative Binomial Exponential

Geometric Gamma

Hypergeometric …
Poisson

…
Foundation Statistics – Discrete Probability Distribution 6
 Bernoulli Distribution
 Discrete probability distribution
 A single experiment with only two mutually exclusive &
collectively exhaustive outcomes: success and failure
(e.g. H & T when tossing a fair coin; Pass & Fail for
driving test; AST or Others when selecting a students of
CCCU)
 X = 1 with success probability p, and 0 with probability of
q=1-p
 Notation: X ~ Bern (p )
 Mean: E (X ) =  = p

 Variance: Var (X ) =  2
= pq

Foundation Statistics – Discrete Probability Distribution 7

 Binomial Distribution: Overview
 Discrete probability distribution
 Repeating Bernoulli trials n times (n represents also the
sample size), with two possible outcomes, success and
failure, each time
 Each trial is independent to others and from identical
setting
– Independent: the result of any trial cannot influence the
results of any other trial
– Identical: each trial must be held under the same conditions
as all other trials exactly, and hence comes with a constant
probability of success p (probability of failure: q or 1-p)
 X : the number of success in the n trials
 Notation: X ~ Bin (n , p ) or X ~ B (n , p )
»
Foundation Statistics – Discrete Probability Distribution 8
 Binomial Distribution: Probability
 If we let 1 = success, & 0 = failure, then a particular
order of X (the no. of success) can be represented as:
111111...1111 + 000000...0000

x successes + (n-x) failures in the n trials

 The joint probability of this particular sequence of
successes & failures can be found as follows
 P(11...100...0)
= p(1) x p(1) x ... x p(1) x p(0) x p(0) x ... x p(0)
= p x p x ... x p x q x q x ... x q
= px qn-x
recall that all trials are independent

Foundation Statistics – Discrete Probability Distribution 9

 Possible Applications for the Binomial
Distribution
 A manufacturing plant labels items as either
defective or acceptable
 A firm bidding for contracts will either get a contract
or not
 A marketing research firm receives survey responses
of “yes I will buy” or “no I will not”
 New job applicants either accept the offer or reject it

Foundation Statistics – Discrete Probability Distribution

 Counting Techniques: Rule of
Combinations
 The number of combinations of selecting x
objects out of n objects is

n!
n Cx =
x!(n − x)!
where:
n! =(n)(n - 1)(n - 2) ... (2)(1)
x! = (X)(X - 1)(X - 2) ... (2)(1)
0! = 1 (by definition)

Foundation Statistics – Discrete Probability Distribution

 Counting Techniques: Rule of Combinations

 How many possible 3 scoop combinations could you

create at an ice cream parlor if you have 31 flavors to
select from and no flavor can be used more than once
in the 3 scoops?
 The total choices is n = 31, and we select X = 3.

31! 31! 31∗ 30 ∗ 29 ∗ 28!

31C3 = 3!(31 − 3)! = 3!28! = 3 ∗ 2 ∗ 1 ∗ 28! = 4,495

Foundation Statistics – Discrete Probability Distribution

 Binomial Distribution: Formula
 Since there are different sequences of 1 & 0, we need to
multiply a number to px qn-x if we want to find probability
of x successes in n trials 11…100…0
00…011…1
10…10…10
01…01…01
and more…

 Combinations of n objects taken x at a time = nCx,

therefore, the formula for the Binomial Distribution
P (X = x )=n C x p x q n −x
For n = 1, 2, 3, ...
x = 0, 1, 2, ..., n

Foundation Statistics – Discrete Probability Distribution 13

 Example: Tossing a fair die
 Example:
Roll a fair die 5 times, find the probability of getting two
“6”

 Answer:
Let X be the number of “6” drawn.
Roll a dice, P(“6”) = 1/6, P(not “6”) = 5/6
Success → “6” { p = 1/6 → q = 1 – 1/6 = 5/6 }
n = 5, X  Bin(5 , 1/6) → P(X) = 5Cx (1/6)x (5/6)n-x
P(two “6”) = P(X = 2) = 5C2(1/6)2(5/6)5-2 = 0.1608

Foundation Statistics – Discrete Probability Distribution 14

 Binomial Distribution: Necessary Conditions
1. Each observation can be classified as one of two mutually
exclusive events
– e.g., success/failure, head/tail, work/not work, study/not study
2. The result of any observation is independent to the result
of any other observations

3. The trials are from the identical population, i.e. the

probability for the two possible outcomes must be
constant from trial to trial
– Each trial should be selected from either:
a finite population with replacement
or an infinite population (with or without replacement)

Foundation Statistics – Discrete Probability Distribution 15

 Example: Drawing balls from a box
 Example:
A box contains four red balls & six black
balls. Three balls are drawn from the
box with replacement. Find the probability
that two red balls are selected.

 Answer: success failure

Checking conditions:
– Two mutually exclusive outcomes only: red / black
– Independent trials: selection of a ball would not affect the next
– Constant probability of success: p = P(red) = 4/10
From the above, X, the no. of red balls selected, follows
Binomial distribution, X ~ Bin(3, 0.4)
Hence, P (X = 2 )= 3 C 2 0.4 0.6 = 0.2880
2 3−2

Foundation Statistics – Discrete Probability Distribution 16

 Example: Computer Shock Test
 Example:
The probability that a brand of notebook computer will
survive a shock test is 0.75.
Find the probability that only 2 of the next 6 computers
tested survive?

 Answer:
Let X be the no. of survived computers after the shock
test, X ~ Bin(6, 0.75).
P(X=2) = 6C2 (0.75)2 (0.25)6-2 = 0.0330

Foundation Statistics – Discrete Probability Distribution 17

 Example: Computer Shock Test (cont’d)
 Continue from the previous example, Find the probability
that
a) Less than 2 of the next 6 computers tested survive;
b) No more than 2 of the next 6 computers tested survive;
c) More than 2 of the next 6 computers tested survive.
 Answer:
a) P(X < 2) = P(X=0) + P(X=1)
= 6C0 (0.75)0 (0.25)6-0 + 6C1 (0.75)1 (0.25)6-1
= 0.0002 + 0.0044 = 0.0046
b) P(X ≤ 2) = P(X=0) + P(X=1) + P(X=2) = 0.0376
c) P(X > 2) = P(X=3) + P(X=4) + P(X=5) + P(X=6)
= 0.9624
Alternatively, P(X > 2) = 1 – P(X ≤ 2) = 1 – 0.0376
=0.9624
Foundation Statistics – Discrete Probability Distribution 18
 Binomial Distribution: Expectation and Variation
 Mean: E (X ) =  = np
 Variance: Var (X ) =  2 = npq
 Standard Deviation: SD (X ) =  = npq

 Revisit the example of computer shock test,

E (X ) =  = np = 6  0.75 = 4.5
Var (X ) =  2 = npq = 6  0.75  0.25 = 1.125
SD (X ) =  = npq = 1.125 = 1.061

Foundation Statistics – Discrete Probability Distribution 19

 In-class Exercise: AD Graduates
 According to past history, the probability that an AD
graduate gets an offer of UGC-funded program is 0.3.
This year, ten selected fresh graduates from the college
apply for UGC-funded program.
a) Find the probability that four graduates get offers?
b) Find the probability that there are four to six graduates
getting offers?
c) Suppose the college awards a scholarship of $5000 for each
fresh graduate getting offer of UGC-funded program.
Find the mean and the standard deviation of the amount of
“UGC scholarship” spent this year.

Foundation Statistics – Discrete Probability Distribution 20

 In-class Exercise: AD Graduates (cont’d)
 Answer

Foundation Statistics – Discrete Probability Distribution 21

 In-class Exercise: AD Graduates (cont’d)
 Answer

Foundation Statistics – Discrete Probability Distribution 22

 Negative Binomial Distribution
 Discrete probability distribution
 Similar to Binomial distribution (independent Bernoulli
trials with two possible outcomes and constant probability
of success), except that the experiment of repeated trials
continues until r successes has occurred
 X : the number of the trial at which the rth success occurs
 Formula: P (X = x | p , r )= C p r q x −r
x −1 r −1

Mean: E (X ) = r
 p
Variance: Var (X ) = rq
p2

Foundation Statistics – Discrete Probability Distribution 23

 Example: Basketball Free Flow
 Example:
Donald is a basketball player, and makes a free throw
70% of the time (i.e. the probability of making a free
throw for Donald is 0.70). During the season, what is the
probability that Donald makes his third free throw on his
fifth shot?
 Answer:
Let X be the no. of free flow required until
Donald makes three shots.
By checking the conditions, X follows
negative binomial distribution.
The required probability is given as:
P (X = 5 | 0.7,3)=5−1 C 3−1 0.7 30.35−3 = 0.1852

Foundation Statistics – Discrete Probability Distribution 24

 Geometric Distribution
 Discrete probability distribution
 A special case of negative binomial distribution,
concerning the number of trials required to get one
success
 X : the number of the trials required to get one success
 Notation: X ~ Geom (p )
 Formula: P (X = x ) = pq x −1
Mean: E (X ) = 1
 p

 Variance: Var (X ) = q
p2

Foundation Statistics – Discrete Probability Distribution 25

 Example: Basketball Free Flow (Revisit)
 Revisit the basketball free flow example,
what is the probability that Donald makes hit first free
throw on his fifth shoot?
 Answer:
Let X be the no. of free flow required until Donald makes
his first shot, X ~ Geom(0.7).
 P (X = 5) = 0.710.35−1 = 0.0057
Alternatively, we can consider X as a negative binomial
random variable, and get
P (X = 5 | 0.7,1)=5−1 C 1−1 0.710.35−1 = 0.0057
which yields the same answer.

Foundation Statistics – Discrete Probability Distribution

 Hypergeometric Distribution: Overview
 Discrete probability distribution
 A sample of size n is randomly selected without
replacement from a finite population of N items.
The trials are hence dependent and the probability of
success is not constant

 Two possible outcomes for each trial: successes and

failures

 In the population, k items can be classified as successes,

and N - k items can be classified as failures
 X : the number of success in the n trials

Foundation Statistics – Discrete Probability Distribution 27

 Hypergeometric Distribution: Formula
 Notation: X ~ Hyper (n , k , N )
Formula: C x N −k C n −x
P (X = x ) =

k

NCn

nk
 Mean: E (X ) =
N

 Variance: Var (X ) = nk (N − k )  N − n
N2 N −1

Foundation Statistics – Discrete Probability Distribution 28

 Example: Drawing balls from a box (Revisit)
 Example:
A box contains four red balls & six black
balls. Five balls are drawn from the
box without replacement. Find the probability
that
a) three red balls are selected.
b) more than four red balls are selected.
 Answer:
Let X be the no. of red balls selected. X follows
Hypergeometric distribution and X ~ Hyper(5,4,10)
C x N −k Cn − x 4 C3 10 − 4 C5−3
a) P ( X = 3) = k
= = 0.2381
N Cn 10 C5

b) P ( X  4) = 0 (why?)

Foundation Statistics – Discrete Probability Distribution 29

 Poisson Distribution: Overview
 Discrete probability distribution
 Used to determine the probability of a certain occurrence
(success) in a fixed interval of time/ length/ area/ volume,
etc.
 The probability of occurrence is fixed
 The number of events occurring in disjoint intervals are
independent
 X : the number of occurrences in a fixed interval
e.g. the number of incoming phone calls per hour for a service center
the number of customer arrivals per day
the number of typos per page of an essay
the number of bacteria per cm3 of water

Foundation Statistics – Discrete Probability Distribution 30

 Poisson Distribution: Formula
 Notation: X ~ Po ( )
where  (lambda), the only parameter necessary to
characterize a Poisson population, is the average
number of occurrences
Formula: x
P (X = x ) = e −


x!
e is the base of the natural logarithm system (2.71828...)

 Mean: E (X ) = 
 Variance: Var (X ) = 

Foundation Statistics – Discrete Probability Distribution 31

 Example: Traffic Accidents
 Example:
The average of traffic accidents on a
highway is two per week. Assume that
the number of accidents follows a
Poisson Distribution.
Find the probability that
a) no accident on this highway during a one-week period;
b) at most two accidents on this highway during a one-
week period;
c) at least two accidents on this highway during a one-
week period;
Find also
d) the standard deviation of the number of accidents on
this highway during a one-week period.
Foundation Statistics – Discrete Probability Distribution 32
 Example: Traffic Accidents (cont’d)
 Answer:
Let X be the number of accidents on this highway during
a one-week period, X0 ~ Po(2).
a) P (X = 0 ) = e −2  2 = 0.1353
0!
b) P (X  2) = P (X = 0 ) + P (X = 1) + P (X = 2)
0 1 2
2 2 2
= e −2  + e −2  + e −2  = 0.6767
0! 1! 2!
c) P (X  2) = 1 − P (X  2) = 1 − P (X = 0 ) + P (X = 1)
= 1 − 0.1353 − 0.2707 = 0.5940
d)
 2 =    =  = 2 = 1.4142

Foundation Statistics – Discrete Probability Distribution 33

 Poisson Distribution: Sum of Variables
 Sum of two independent Poisson distributed random
variables come up with a Poisson distribution

 If X and Y are two independent Poisson distributed

random variables, with mean X and Y, then
X +Y ~ Po (X + Y )

Foundation Statistics – Discrete Probability Distribution 34

 In-class Exercise: Game Console
 Suppose that the no. of customer orders for a latest game
console is distributed with a Poisson distribution with the
average number of order per hour as 3.
What is the probability of getting twenty orders by the
end of an 8-hour working day?

Foundation Statistics – Discrete Probability Distribution 35

 In-class Exercise: Game Console (cont’d)
 Answer:

Foundation Statistics – Discrete Probability Distribution 36