Recall Lec 4 know when to use PMF or PDF

Common Discrete Distributions

Math 150.1

Institute of Mathematics
University of the Philippines Diliman

1/ 29

Math 150.1 Common Discrete Distributions Lecture 5 1 / 29

Discrete Distributions
Recall: A random variable X is called a discrete random variable if its
range is countable.

1 Uniform / Empirical Distribution

2 Bernoulli and Binomial Distributions

3 Poisson Distribution

4 Hypergeometric Distribution

5 Geometric and Negative Binomial Distributions

2/ 29

Math 150.1 Common Discrete Distributions Lecture 5 2 / 29

Uniform Distribution

X N UnifCA
Definition
Let X be a discrete random variable and A be a finite set. We say
that X has a uniform distribution on A, denoted by X ⇠ U(A), if
the probability on each number in A is equal. That is, the pmf of X
is defined by
1
f (x) = P[X = x] = A (x),
|A|
indicator that x
where |A| is the cardinality of A. should be in A

3/ 29

Math 150.1 Common Discrete Distributions Lecture 5 3 / 29

Uniform Distribution
Theorem
If X ⇠ U{1, . . . , m} then
m+1
I X EI K P X K
1 the mean is µ = ,
2
2
2 m 1
2 the variance is = , and
12
3 the momemt generating function for X is X
EI KZ.PE
e t (1 e tm ) L Em K2
MX (t) = t
.
m(1 e )
Nimtflamtd
02 11 x2 m2
Proof: (exercise)
2mztzmt mztgmtt e.ms
4/ 29

Math 150.1 Common Discrete Distributions Lecture 5 4 / 29

Uniform Distribution
Suppose X ⇠ U{1, . . . , m} and if we let Y = X 1, then Y is a
discrete r.v. with Y ⇠ U{0, . . . m 1}. The pdf of Y is given by
1
f (y ) = where y = 0, 1, 2, . . . , m 1. In this case,
m
m+1
µY = E [Y ] = E [X 1] = E [X ] 1= 1
2
m 1
=
2
2
2 m 1
Y = Var [Y ] = Var [X 1] = Var [X ] =
12
MY (t) = E [e Yt ] = E [e Xt t ] = e t E [e Xt ] = e t MX (t)
1 e tm no et since different set
=
m(1 e t )
5/ 29

Math 150.1 Common Discrete Distributions Lecture 5 5 / 29

Exercises
1 A six-faced fair die is rolled. Suppose X denote the number appear
on the top of a die.
a. Find the probability that an even number appear on the top.
b. Find the probability that the number appear on the top is less
than 3.
c. Compute mean and variance of X
2 A telephone number (7 digits) is selected at random from a directory.
Find the probability that the last digit of the selected number is (a) a
6; (b) less than 3; (c) greater than or equal to 8.
3 Let X be an integer chosen at random from the interval [a, b] where a
and b are integers. Find the probability mass function f (x), its mean
and its variance.
4 Suppose X ⇠ U{1, 2, 5, 7, 8}. Find the mean and variance of X .
6/ 29

Math 150.1 Common Discrete Distributions Lecture 5 6 / 29

 19 IP Ix is even
3 I
b IP x 3 P x IP X 2

t
E

CI Mx
GI
o m distribution
E the oven
O
off 3g

P
I I I 101 to Pix 6 28
iI
PC x 3
1Mx 9
X N M O I 213 T9

Take 2 Uft it a ti
bl b
3 X n u a a ti at2 Z a t t
nI l a att b i by b at X

gIF
bi b
II x e a ati
a
n
E a a
E.EE
02 Var X
É K2 IP x k MZ

M ELx 4
t 2
t t t8
t I 123
IE x t a t t 64 f

O II Li M
Bernoulli Distribution
When one tries to perform a single experiment with only two possible
outcomes, success (1) or failure (0), then what we have is a Bernoulli
distribution with a probability of success equal to p.
x Beep failure e e
x son
acres p

Definition
A random variable X follows a Bernoulli distribution, denoted by
X ⇠ Be(p), if the pmf of X is given by

f (x) = p x q 1 x
{0,1} (x)

where p 2 [0, 1] and q = 1 p.

7/ 29

Math 150.1 Common Discrete Distributions Lecture 5 7 / 29

Bernoulli Distribution

Theorem
If X ⇠ Be(p), then

E(X ) = p, Var (X ) = pq and MX (t) = pe t + q.

M 7 Pax 1 OCP x07 E Ex p
I p toed vary p p pq
Proof: (Left as an Exercise) p

Exercise: In a toss of a biased coin that is twice as likely to come up

heads as tails, what is the probability of getting a head? a tail?

8/ 29

Math 150.1 Common Discrete Distributions Lecture 5 8 / 29

Binomial Distribution
Suppose the experiment is done n times. Then the experiment
becomes a binomial experiment and the rv X that counts the number
of successes (with probability p) out of the n trials.

Definition
A random variable X follows a binomial distribution, denoted by
X ⇠ Bi(n, p), if the pmf of X is given by X Bill D X BEE
✓ ◆
n x n x
f (x) = p q {0,1,...,n} (x)
x

where n is a positive integer, p 2 [0, 1] and q = 1 p.

9/ 29

Math 150.1 Common Discrete Distributions Lecture 5 9 / 29

Binomial Distribution

Theorem
IF X ⇠ Bi(n, p), then

E(X ) = np, Var (X ) = npq and MX (t) = (pe t + q)n .

Proof: (Exercise) Formalexercise

10/ 29

Math 150.1 Common Discrete Distributions Lecture 5 10 / 29

Exercises

1 If we toss a biased coin that is twice as likely to come up heads

as tails, five times, what is the probability of getting 2 heads out
of the 5 tosses? at least 2? at most 2?
2 Find the probability that in five tosses of a fair die a 3 appears
(a) twice, (b) at most once, (c) at least two times?
3 Find the probability that in a family of 4 children, there will be
(a) at least 1 boy, (b) at least one boy and at least one girl.
Assume that a boy and a girl have the same probability of
getting born.

11/ 29

Math 150.1 Common Discrete Distributions Lecture 5 11 / 29

I X Bi 5 2
3 success getting H ID1 22 P x 2 P x 3

P1 2 E 3 b
4
MI
401243
PI XI 2 I
AtFy

PATH Hats
2 IP 3
I
Bi 1516
P P x
371
27 11211112

Y eD
x Pixel
Y

1 b IP 11 13
3 IP boy 2 L IP X O IPA 4
x Bi 4 12 4
9 P x2 1 l P x0 1 1811474 4
trinity
Poisson Distribution

Suppose we would like to know how many times a person sends a

text message within an hour. One can model the number of
occurrences per unit time using a Poisson Distribution.

Definition
A random variable X follows a Poisson distribution with parameter
> 0, denoted by X ⇠ Poi( ), if the pmf of X is given by
IMPORTANT
Arithmeticseriestormu ia e x

Ifatin expansion f (x) = {0,1,2,...} (x)

of e x!
Remark

É IP xx l

E
e e I 12/ 29

Math 150.1 Common Discrete Distributions Lecture 5 12 / 29

Poisson Distribution

Theorem
If X ⇠ Poi( ), then

E(X ) = , Var (X ) = and MX (t) = exp[ (e t 1)].

Proof: (Exercise) Part of activity usingsummations

13/ 29

Math 150.1 Common Discrete Distributions Lecture 5 13 / 29

Exercises

1 If X has a Poisson distribution with P[X = 1] = P[X = 2], what

is (a) E[X ]; (b) P[X = 1 or 2]?
2 The moment generating function of a random variable X is
4(e t 1)
e . Show that P[µ 2 < X < µ + 2 ] = 0.931.
3 Let X be a Poisson distribution with E[X ] = 100. Use
Chebychev’s Inequality to determine a lower bound for
P[75 < X < 125].

14/ 29

Math 150.1 Common Discrete Distributions Lecture 5 14 / 29

I X n Poi X
p x1
st Pix 1 P x2
P x 2 e
M E X
Ky
D 1 2 2 IPxen

ZX X dn
legz.at 21 HI

2 M It e4 et i
x poi14 M 4and 0 2 IP M 20examtee
mf
Wecantindelxf 14
land
exam.cn 4
IP lx MI ro s're usingthedistributionofX

Pl ro x micro É PIMLOLXLUTZE
IP 05 98
PCu roc X M ro f
IP x 1 P x 2 t
IPA 7
e 4 ate te t
t
HI
Binomial-Poisson Approximation

Remark: In the binomial distribution, if n is sufficiently large (in

practice n 50) and p is sufficiently small (such that np < 5) so
that q is very close to 1, the event is called a rare event. When this
happens, the binomial distribution approximates the Poisson
distribution so that = np, q ⇡ 1. We formalize this remark in the
following theorem.

Theorem
If n is large and p is close to 0, then the binomial distribution
approximates the Poisson distribution with = np.

15/ 29

Math 150.1 Common Discrete Distributions Lecture 5 15 / 29

The Poisson Distribution
✓ ◆
n x n x
Proof: Since X ⇠ Bi(n, p) then f (x) = p q where
x
E(X ) = np. Let = np so that p = /n. Then
✓ ◆ ✓ ◆x ✓ ◆n x
n
f (x) = 1
x n n
✓ ◆n x
n(n 1)(n 2) · · · (n x + 1) x
= 1
x!nx n
1 ✓ ◆n
1 1 n
1 n2 · · · 1 x n 1 x
x
= 1
x! n

16/ 29

Math 150.1 Common Discrete Distributions Lecture 5 16 / 29

Poisson Distribution

Then
1 2 x 1 ✓ ◆n x
1 1 n
1 n
··· 1 n x
f (x) = 1
x! n

As n approaches infinity, the numerator of the first factor approaches

1, while the last factor approaches to e . Thus, if we let n approach
infinity we have
x
e
f (x) ⇡
x!
which is the pmf of a Poisson distribution. QED

17/ 29

Math 150.1 Common Discrete Distributions Lecture 5 17 / 29

Example

If the probability that an individual suffers a bad reaction from

injection of a given serum is 0.0001, determine the probability that
out of 2 000 individuals (a) exactly 3, (b) more than 2, individuals
will suffer a bad reaction.

Solution: Let X represent the random variable of the number of

individuals who will suffer a bad reaction. From the given problem,
n = 2, 000 and p = 0.0001. Note that n is large enough and p is
small enough that we can let = 0.2. By the previous theorem,

X ⇠ Bi(2 000, 0.0001) ! X ⇠ Poi(0.2)

18/ 29

Math 150.1 Common Discrete Distributions Lecture 5 18 / 29

X Bi 2000 0 001
a IP x 3 20300 0 0017310.999971997

X I Y where Y Poi i O 2

IP x 3 P Y 3
093,10271
b IP X I 2 I IPLX O IP X D

but
IP IX 2 I 1171422
I P 14 0 IP Y
The Hypergeometric Distribution

Consider a random experiment consisting of selecting n objects

without replacement out of M dichotomous objects. Out of the M
objects, there are K of one kind (“success”) and there are M K of
another kind (“failure”). Let X be the number of successes out of the
n selected elements. Then X follows a hypergeometric
distribution.
Definition
X follows a hypergeometric distribution, denoted by nobiems
iffy
onieces
norep
X ⇠ HG (n, M, K ), if its pmf is given by
K M K ÉÉ aim
x n x
f (x) = M {max(0,n+K M),1,...,min(n,K )} (x).
n

19/ 29

Math 150.1 Common Discrete Distributions Lecture 5 19 / 29

The Hypergeometric Distribution

Theorem
If X ⇠ HG (n, M, K ), then

nK nK M K M n
E(X ) = , Var (X ) = .
M M M M 1

Proof: (Left as an Exercise) look for the

avert online
Remark: The moment generating function for a hypergeometric
random variable has no simple closed form.

20/ 29

Math 150.1 Common Discrete Distributions Lecture 5 20 / 29

Exercises

1 There are 10 cards lying face down on a table and 2 of them are
aces. If 5 cards are selected at random, what is the probability
that 2 of them are aces?

2 A baseball team has 16 non-pitchers on its roster. Of these, 6

bat left-handed and 10 right-handed. The manager, having
already selected the pitcher for the game, randomly selects 8
players for the remaining positions.
a. What is the probability that he selects 4 left-handed batters and
4 right-handed batters?
b. What is the expected number of left-handed batters chosen?

21/ 29

Math 150.1 Common Discrete Distributions Lecture 5 21 / 29

112
X n
it 5,10127
ex
I I
a

HG 8 16 6
2 Y
4 47
i
b E EY
Eiji
Geometric and N.B.

In a binomial distribution, the fixed parameters are the number

of trials n and the probability of success p, and X represents the
number of successes out of the n trials.
In a negative binomial distribution, the fixed parameters are
the probability of success p and the number of successes r , and
X represents the number of trials up to the r th success.
A special case of a negative binomial distribution is when r = 1,
in which case, it is commonly known as the geometric
distribution.
A random variable X following a geometric distribution is often
called the discrete waiting-time variable (we looking at how
many trials will it take to get the 1st success).
22/ 29

Math 150.1 Common Discrete Distributions Lecture 5 22 / 29

Geometric Distribution

Definition
Suppose X is a random variable representing the number of trials
until the first success (with probability p), then X follows a
geometric distribution, denoted by X ⇠ Geo(p), with pmf given by
1
f (x) = q x p {1,2,...} (x), where q = 1 p.

Theorem
If X ⇠ Geo(p), then

1 q pe t
E(X ) = ; Var (X ) = 2 and MX (t) = .
p p 1 qe t

23/ 29

Math 150.1 Common Discrete Distributions Lecture 5 23 / 29

Negative Binomial

Definition
Suppose X is a random variable representing the number of trials
until the r th success (with probability p), then X follows a negative
binomial distribution, denoted by X ⇠ NB(r , p), with pmf given by
✓ ◆
x 1 x r r
f (x) = q p {r ,r +1,r +2,...} (x),
r 1

where q = 1 p.

24/ 29

Math 150.1 Common Discrete Distributions Lecture 5 24 / 29

Negative Binomial Distribution

Under this definition, we have the following theorem:

Theorem
If X ⇠ NB(r , p), then
 t r
r rq pe
E(X ) = ; Var (X ) = 2 and mX (t) = .
p p 1 qe t

Proof: (Left as an Exercise)

25/ 29

Math 150.1 Common Discrete Distributions Lecture 5 25 / 29

N.B. Distribution

Alternatively, we may consider the random variable Y as the number

of failures until the r th success, success of which has probability p.
Definition
Suppose Y is a random variable representing the number of failures
until the r th success (success of which has probability p), then Y
follows a negative binomial distribution, denoted by Y ⇠ NB(r , p),
with pmf given by
✓ ◆
r +y 1 y r
f (y ) = q p {0,1,...} (y ),
y

where q = p 1.

26/ 29

Math 150.1 Common Discrete Distributions Lecture 5 26 / 29

N.B. Distribution

Theorem
If Y ⇠ NB(r , p), then
 r
rq rq p
E(Y ) = ; Var (Y ) = 2 and mY (t) = .
p p 1 qe t

Proof: (Left as an Exercise)

27/ 29

Math 150.1 Common Discrete Distributions Lecture 5 27 / 29

Exercises
1 An experiment consists of drawing a card at random from a
standard deck and replacing it. If this experiment is done
repeatedly, what is the probability that
a. the first heart appears on the fifth draw?
b. the first ace appears on the tenth draw?
c. the third heart appears on the tenth draw?

2 At a clinic, patients are given X-rays to test for tuberculosis. If

15% of all the patients have tuberculosis,
a. what is the probability that the 15th patient tested will be the
third with tuberculosis?
b. what is the mean number of patients without tuberculosis tested
before the sixth patient with tuberculosis is tested?

28/ 29

Math 150.1 Common Discrete Distributions Lecture 5 28 / 29

1 a X N BI r P Ya
X G to p 4
or simply
1 5
3 I
b Y Eeo p ti
14 10
Ig
Z N3
c r 3 p ta
PCE 1
14
o
2 H Ai w AND2 r 3 p O E

IP Y 15 142 I t 15 C O 85g

151
tf N NB t 16 0 15

t Ifl 0 I2 i
I and for m

ETY 610.11
0 15

or E N IVE I 66 E 15
E 6 7 firstterm
IE E E I
II III I 6
 P I Io IFI WitID x ID IF I X 121 IP X I3 Tx 14
A
ITCX 15
0.411010 6 Y ii 47 t.at
É
is ii a

IP1.3 E Y E 8

a 1718 5 I 0.415 0 6

9,3
I 1 ai
b Y HimREG n 5 M 25 88 55 I
Men
YEÉ
i
6 X HG n 10 m 20 K 5

1 51 1

9 5000 policies 12 200 claims

10 years I a ve of claims
per policy per year

o 110 1
b Xxx Poi 10.241441
P XLI IE O PIE L
t.tt
e 1o.z44it t e 44
ot44yo.i

te
c I 10001J 1000 FIX
1000 o 44 fI44
 Math 150.1

29/ 29

Math 150.1 Common Discrete Distributions Lecture 5 29 / 29