UNIT II

Some Special Probability Distributions

1 Binomial Distribution
Definition
Let A be some event associated with a random experiment E, such that P(A) = p and P(Ā) = 1− p = q.
Assuming that p remains the same for all repetations, if we consider n independent repetations (or
trials) of E and if the random variable (RV) X denotes the number of times the event A has occured,
then X is called a binomial random varaiable with parameters n and p or we say that X follows a bi-
nomial distribution with parameters n and p, or symbolically B(n, p). Obviously the possible values
that X can take, are 0, 1, 2, ..., n.
By the theorem under Bernoulli’s trials in Chapter 1, the probability mass function of a binomial RV
is given by

P(X = x) = nCr px qn−x ; x = 0, 1, 2, ..., n where p+q = 1

Examples:
1. Tossing coin once results is head or tail.
2. Toss a dice once, the outcomes can be classified as ’even number’ and ’odd number’.
3. Firing at a target results in hit or miss.

1.1 Moment Generating Function (MGF) of Binomial distribution

Binomial distribution is

p(x) = P(X = x) = nCx px qn−x ; x = 0, 1, 2, ..., n

MGF is MX (t) = E[etX ]

= ∑nx=0 etx p(x)

= ∑nx=0 etx nCx px qn−x

= ∑nx=0 nCx(pet )x qn−x

=(q + pet )n

1
Mean and Variance of Binomial distribution
We know that MGF is
MX (t) = (q + pet )n
∴ MX′ (t) = n(q + pet )n−1 .pet
MX′′ (t) = np (n − 1)(q + pet )n−2 (pet ).et + (q + pet )n−1 .et


= np(q + pet )n−2 et (n − 1)pet + q + pet

putting t = 0, we get MX′ (0) = n(q + p)n−1 .p

=⇒ mean = µ1′ = np
MX′′ (0) = np(n − 1)p + q + p
=⇒ µ2′ = np(np + q) = n2 p2 + npq
Var(X) = E(X 2 ) − [E(X)]2 = µ2′ − (µ1′ )2
= n2 p2 + npq − n2 p2 = npq

Problems
1. The mean of a binomial distribution 20 and standard deviation is 4. Find the parameters of the
distribution.
Solution:
Let (n, p) be the parameters of the binomial distribution, then

√
mean = np and S.D= npq
√
Given, np =20 and npq =4
npq =16
∴ 20.q = 16
4
=⇒ q =
5
4 1
∴ p = 1−q = 1− =
5 5
1
∴ n. = 20 =⇒ n = 100
5
1
∴ the parameters are 100, .
5

2.Six dice are thrown 729 times. How many times do you expect atleast 3 dice to show a five or a six?
Solution:
Success is getting 5 or 6 in a die. Let X denote the numer of success when 6 dice are thrown.
∴ X is a binomial R.V with parameter (n, p)

∴ P(X = x) = nCx px qn−x , x = 0, 1, 2, ..., n

2 1
Given n = 6 and p = P(5 or 6) = =
6 3
1 2
∴ q = 1− p = 1− =
3 3
 x  6−x
1 2
∴ P(X = x) = 6Cx , x = 0, 1, 2, ...6
3 3

2
 P(X ≥ 3) = 1 − P(X < 3)
(   2  4 )
2 6 1 2 5
 
1 2
= 1− + 6C1 + 6C2 .
3 3 3 3 3
1
= 1− {64 + 192 + 240}
36
496 233
= 1− =
729 729
233
When 6 dice are thrown 729 times, the number of times atleast 3 dice show 5 or 6 is 729× =233.
729

3.Out of 800 families with 4 children each, how many families would be expected to have (a) 2
boys and 2 girls,(b) atleast 1 boy, (c) at most 2 girls, and (d) children of both sexs. Assume equal
probabilities for boys and girls.
Solution:
1 1
Considering each child as a trial, n = 4. Assuming that birth of a boy is success, p = and q = .
2 2
Let X denote the number os successes (boys)

(a)P(2 boys and 2 girls) =P(X = 2)

 2  4−2
1 1
= 4C2 . .
2 2
 4
1 3
=6 × =
2 8

∴ No.of families having 2 boys and 2 girls

= N.P(X = 2)(where N is the total no.of families considered)

3
= 800 ×
8
= 300.

(b)P(atleast 1 boy) = P(X ≥ 1)

= P(X = 1) + P(X = 2) + P(X = 3) + P(X = 4)
= 1 − P(X = 0)
 0  4
1 1
= 1 − 4C0 . .
2 2
1 15
= 1− =
16 16
∴ No.of families having atleast 1 boy
15
=800 × = 750.
16

3
 (c)P(at most 2 girls) =P(exactly 0 girl, 1 girl or 2 girl)
= P(X = 4, X = 3orX = 2)
= 1 − {P(X = 0) + P(X = 1)}
(  4  4 )
1 1
= 1 − 4C0 . + 4C1 .
2 2
11
=
16
∴ No.of families having atmost 2 girls
11
= 800 × = 550.
16

(d)P(children in both sexs)

= 1 − P(children in same sex)
= 1 − {P(all are boys) + P(all are girls)}
= 1 − {P(X = 4) + P(X = 0)}
(  4  4 )
1 1
= 1 − 4C4 . + 4C0 .
2 2
1 7
= 1− =
8 8
∴ No.of families having children of both sexs
7
= 800 × = 700.
8

4. An irregular 6-faced die is such that the probability that is given 3 even numbers in 5 throws
is twice the probability that it gives 2 even numbers in 5 throws. How many sets of exactly 5 trials
can be expected to give no even number out of 2500 sets?
Solution:
Let the probability of getting an event number with the unfair die be p.
Let X denote the number of even numbers obtained in 5 trials (throws).

P(X = 3) = 2 × P(X = 2)
i.e., 5C3 p3 q2 = 2 × 5C2 p2 q3
i.e., p = 2q = 2(1 − p)
2 1
∴ 3p = 2 orp = and q =
3 3

Now,P(getting no even number)

= P(X = 0)
 5
0 5 1 1
= 5C0 .p .q = =
3 243

∴ Number of sets having no success (even number) out of N sets = N × P(X = 0)

1
∴ Required number of sets = 2500 × = 10, nearly
243
5. Two dice are thrown 120 times. Find the average number of times in which the number of the

4
first dice exceeds the number on the second dice.
Solution:
The number on the first dice exceeds that on the second dice, in the following combinations:
(2,1);(3,1),(3,2);(4,1),(4,2),(4,3);(5,1),(5,2),(5,3),(5,4);(6,1);(6,2),(6,3),(6,4),(6,5),
where the numbers in the parentheses represent the numbers in the first and second dice respectively.
∴ P(success) = P(number in the first dice exceeds the number in the second dice)

15 5
= =
36 12
This probability remains the same in all the throws that are independent.
If X is the number of success, then X follows a binomial distributions with parameters n(= 120) and
5
p(= ).
12
5
∴ E(X) = np=120 × = 50
12
6. Fit a binomial distribution for the following data:

x: 0 1 2 3 4 5 6 Total
f: 5 18 28 12 7 6 4 80
Fitting a binomial distribution means assuming that the given distribution is approximately binomial
and hence finding the probability mass function and then finding the theoretical frequencies.
To find the binomial frequency distribution N(q + p)n , which fits the given data, we required N, n and
p. We assume N=total frequency = 80 and n= no. of trials =6 from the given data.
To find p, we compute the mean of the given frequency distribution and equate it to np (means of the
binomial distribution).

x: 0 1 2 3 4 5 6 Total
f: 5 18 28 12 7 6 4 80
fx: 0 18 56 36 28 30 24 192

∑ f x 192
x= = = 2.4
∑f 80
i.e., np = 2.4or 6p = 2.4
∴ p = 0.4and q = 0.6

If the given distribution is nearly binomial, the theoretical frequencies are given by the successive
terms in the expansion of 80(0.6 + 0.4)6 . Thus we get,

x: 0 1 2 3 4 5 6
Theoretical f : 3.73 14.93 24.88 22.12 11.06 2.95 0.33
Converting these values into whole numbers consistent with the condition that the total frequency is
80, the corresponding binomial frequency distribution is as follows:

x: 0 1 2 3 4 5 6 Total
f : 4 15 25 22 11 3 0 80

5
2 Poisson Distribution
Definition
If X is a discrete RV that can assume the values 0, 1, 2, ..., such that its probability mass function is
given by

e−λ λ x
P(X = x) = ; x = 0, 1, 2, ...; λ >0
x!
then X is said to follow a Poisson distribution with parameter λ or symbolicaly X is said to follow
P(λ ).

Poisson Distribution on Limiting form of Binomial Distribution

Poisson distribution is a limiting case of binomial distribution under the following conditions:

1. n, the number of trials is indefinitely large, i.e., n → ∞,

2. p, the constant probability of success in each trial is very small, i.e., p → 0.

λ λ
3. np(= λ ) is finite or and q = 1 − , where λ is a positive real number.
n n

Examples
1. The number of defective items produced in a factory in one day.
2. The number of deaths due to rare disease.
3. The no. of mistakes in a page of a book.
4. The number of earthquakes in a given time interval.

2.1 Moment Generating Function of Poisson distribution

Let X be a Poisson randon variable with parameter λ
e−λ λ x
∴ Poisson distribution is given by p(x) = P(X = x) = , x = 0, 1, 2, ...
x!
Moment generating function of X is

MX (t) = E[etX ] = ∑∞ tx
x=0 e p(x)

tx e−λ λ x
= ∑∞
x=0 e
x!
t x
−λ ∞ (λ e )
= e ∑x=0
x!
t
= e−λ .eλ e
t
MX (t) = eλ (e −1)

6
Mean and Variance of Poisson distribution
We have
t
MX (t) = eλ (e −1)
t
∴ MX′ (t) = eλ (e −1) .λ et
n t t
o
MX′′ (t) = λ eλ (e −1) .et + et .eλ (e −1) .λ et
t
= λ eλ (e −1) .et {1 + λ et }
putting t = 0, MX′ (0) = λ =⇒ µ1′ = λ
MX′′ (0) = λ (1 + λ ) =⇒ µ2′ = λ (1 + λ )
∴ Mean = µ1′ = λ
Var(X) = µ2′ − (µ1′ )2
= λ +λ2 −λ2

Note: The mean and variance of Poisoon distribution are equal.

Problems
1. If X is a poisson R.V such that P(X = 2) = 9 P(X = 4) + 90 P(X = 6), then find (i) the variance,
(ii) the mean , E(X 2 ), P(X ≥ 2).
Solution:

e−λ λ x
Poisson distribution is P(X = x) = , x = 0, 1, 2, ...
x!
Given P(X = 2) = 9 P(X = 4) + 90 P(X = 6)
λ2 e−λ λ 4 e−λ λ 6
=⇒ e−λ = 9. + 90.
2! 4! 6!
1 3 2 1 4
=⇒ = λ + λ [Dividing bye−λ λ 2 ]
2 8 8
=⇒ λ 4 + 3λ 2 − 4 = 0
=⇒ (λ 2 + 4)(λ 2 − 1) = 0
=⇒ λ = 1 [∴ λ > 0]
We know that Mean= λ = 1; Var(X)λ = 1
E(X 2 ) = λ 2 + λ = 1 + 1 = 2
P(X ≥ 2) = 1 − P(X < 2)
e−1 .1
 
= 1 − [P(X = 0) + P(1)] = 1 − e + −1 = 1 − 2e−1
1!

2. Find the probability that at most 4 defective fuses will be found in a box of 200 fuses if experience
shows that 2% of such fuses are defective.(given e−4 = 0.0183).
Solution:

Let X denote the number of defectives in a box of 200 fuses then

p = probability of defective fuses.
2
= = 0.012
100

7
 Given n = 200 and p is very small.
∴ X is a Poisson R.V with parameter λ
2
∴ λ = np = 200 × = 4.
100
e−λ λ x
So, the poisson distribution is P(X = x) = , x = 0, 1, 2, ...
x!
e−4 4x
= , x = 0, 1, 2, 3, ...
x!
Required P(X ≤ 4) = P(X = 0) + P(X = 1) + P(X = 2) + P(X = 3) + P(X = 4)
4 42 43 44
= e−4 + e−4 . + e−4 . + e−4 . + e−4 .
 1! 2!  3! 4!
32 32
= e−4 1 + 4 + 8 + +
3 3
 
64
= e−4 13 +
3
103
= 0.0183 × = 0.628
3

3.The number of accidents in a year attributed to taxi drivers in a city follows a poisson distribution
with mean equal to3. Out of 1000 taxi drivers, find approximately the number of drivers with (i) no
accident in a year, (ii) more than 3 accidents in a year.
Solution:
Let X denote the number of accidents in a yaer due to taxi drivers.
Given X is a Poisson distribution. Let λ be the parameter.
Given mean =3 =⇒ λ = 3

e−λ λ x
∴ Poisson distribution is given by P(X = x) = , x = 0, 1, 2, ...
x!
e−3 .3x
= , x = 0, 1, 2, ...
x!
(i)P(X = 0) = e−3 = 0.0498
∴ number of drivers with no accidents =100 × 0.0498 = 49.8
i.e., number of drivers with no accidents =50.

(ii) P(X > 3) = 1 − P(X ≤ 3)

= 1 − {P(X = 0) + P(X = 1) + P(X = 2) + P(X = 3)}
32 33
 
= 1−e −3 1+3+ +
2! 3!
=1-0.0498 ×13
=1-0.6474=0.3526
∴ number of drivers with more than 3 accidents
=1000 × 0.3526=353.

4.The number of monthly breakdown of a computer is a random variable having a poisson distribution
with mean equal to 1.8. Find the probability that this computer will function for a month
(i)without breakdown

8
(ii) with only one breakdown and
(iii)with atleast one breakdown
Solution:
Let X denote the number of breakdowns of a computer in a month.
Given X follows Poisson distribution with mean =λ =1.8.
∴ Poisson distribution is given by

e−λ λ x
P(X = x) = , x = 0, 1, 2, ...
x!
e−1.8 (1.8)x
= , x = 0, 1, 2, ...
x!

(1)P(X = 0) = e−1.8 = 0.1653

(2)P(X = 1) = e−1.8 (1.8) = 0.2975
(3)P(X ≥ 1) = 1 − P(X = 0) = 1 − 0.1653 = 0.8347

5. Fit a poisson distribution to the data:

x: 0 1 2 3 4 5
f: 142 156 69 27 5 1

Solution:
Poisson distribution

e−λ λ x
P(X = x) = , x = 0, 1, 2, ...
x!
∑ fx
Mean=λ =
∑f

x: 0 1 2 3 4 5 Total
f: 142 156 69 27 5 1 400
fx: 0 156 138 81 20 5 400

∑ f x 400
∴ Mean = = =1
∑ f 400
λ =1
Fitting of Poisson distribution:

x: 0 1 2 3 4 5
observed frequency: 142 156 69 27 5 1
Expected frequency: 147.15 147.15 73.58 24.53 6.13 1.23
Expected(Approx): 147 147 74 25 6 1

3 Geometric Distribution
Definition
Let the RV X denote the number of trials of a random experiment required to obtain the first success
(occurence of an event A).Obviously, X can assume the values 1, 2, 3, ...

9
Now, X = x, if and only if the first (x − 1) trials result in failure(occurance of A ) and the xth trial
results in success (occurrence A). Hence,

P(X = x) = qx−1 p; x = 1, 2, 3, ..., ∞

where P(A) = p and P(A) = q.

If X is a discrete RV that can assume the values 1, 2, 3, ..., ∞ such that its probability mass function is
given by

P(X = x) = qx−1 p; x = 1, 2, 3, ..., ∞ where p + q = 1

then X is said to follow a geometric distribution.

3.1 Moment generating Function of geometric distribution

The geometric distribution with parameter p is

P(X = x) = qx−1 p, x = 1, 2, 3, ...

MX (t) = E(etX )

= ∑∞ tx
x=1 e p(x)

= ∑∞ tx x−1 p
x=1 e q

= ∑∞ t t(x−1) qx−1 p
x=1 e .e

= pet ∑∞ t x−1
x=1 (qe )

= pet [1 + (qet ) + (qet )2 + ...∞]

1
pet if qet < 1
1 − qet
pet
=
1 − qet

10
Mean and variance
pet
MX (t) =
1 − qet
(1 − qet )et − et (−qet )
 
′
∴ MX (t) = p
(1 − qet )2
1 − qet + qet pet
 
= pe t =
(1 − qet )2 (1 − qet )2
(1 − qet )2 .et − et .2(1 − qet )(−qet )
 
′′
Now MX (t) = p
(1 − qet )4

1 − qe t + 2qet 
= pet (1 − qet )
(1 − qet )4
pet (1 + qet )
=
(1 − qet )3
p p 1 1
∴ MX′ (0) = 2
= 2 = =⇒ mean =
(1 − q) p p p
p(1 + q) p(1 + q)
MX′′ (0) = =
(1 − q)3 p3
1 q
=⇒ µ2′ = 2
+ 2
p p

Var(X) = µ2′ − (µ1′ )2

1 q 1 q
= 2
+ 2− 2= 2
p p p p

Property of Geometric Distribution:

[Geometric Random Variable has memoryless property]
Definition:
A non-negative random variable X is memoryless if
P(X > s + t/X > s) = P(X > t)for all s,t ≥ 0.
Proof: The geometric distribution is

P(X = x) = qx−1 p, x = 1, 2, 3, ...

P(X > m) = P(X = m + 1) + P(X = m + 2) + ...∞
= qm p + qm+1 p + qm+2 p + ...∞
= qm p[1 + q + q2 + ... + ∞]
qm p qm p
= = = qm
1− p p
P(X > s + t ∩ X > s)
∴ P(X > s + t/X > s) =
P(X > s)

11
 Since the event {X > s + t} ⊂ {X > s} ,
{X > s + t} ∩ {X > s} = {X > s + t}
P(X > s + t)
∴ P(X > s + t/X > s) =
P(X > s)
qs+t
= = qt = P(X > t)
qs
Note:This property is also known as Markov property and ageless property.

Problems
1.If the propability that a target is destroyed by anyone shot is 0.6 what is the probability that it should
be destroyed on the 5th attempt?
Solution:
Let X denote the number of attempts required for the first success (success is distribution)
Then X follows geometric distribution

P(X = x) = qx−1 p, x = 1, 2, ...

Given p = 0.6 and q = 1 − p = 0.4

Here first success is the 5th attempt
∴x=5

Required P(X = 5) = (0.4)4 (0.6) = 0.01536

2. In a particular manufacturing process it is found that, on the average, 1% of the items is de-
fective. What is the probability that the fifth item inspected is the first defective item?
Solution:
Here defective is success. First defective is the 5th one.
1
So, geometric distribution with p = = 0.01, x = 5.
100
P(X = x) = qx−1 p, x = 1, 2, 3, ...
 4
99 1
∴ P(X = 5) = . = 0.0096.
100 100

3. A die is thrown until 6 appears. What is the probability that it must be thrown more than four
times?
Solution:
Getting 6 is success. For the first success it should be thrown more than 4 times. So, geometric distri-
bution is used.

P(X = x) = qx−1 p, x = 1, 2, ...

1 5
Here p = probability of 6 = ∴ q = 1− p =
6 6

12
Required

P(X > 4) = 1 − P(X ≤ 4)

= 1 − {P(X = 1) + P(X = 2) + P(X = 3) + P(X = 4)}

= 1 − p + qp + q2 p + q3 p


p(1 − q4 ) a(1 − rn )
= 1− [sum of G.P = ]
1−q 1−r

= 1 − (1 − q4 ) = r4

 4
5 625
= = = 0.48.
6 1296

4 Uniform or Rectangular Distribution

Definiton
A continuous RV X is said to follow a uniform or rectangular distribution in any finite interval, if its
probability density function is a constant in that interval.
1
If X follows a uniform distribution in a < x < b, then its pdf is given by f (x) = , a < x < b.
b−a

Moment generating function, Mean and Variance

Z b
tX
MX (t) = E(e ) = etx f (x)dx
a
Z b
1
= etx dx
a b−a
 tx b
1 e 1
= = (etb − eta ),t ̸= 0
b − a t a t(b − a)
1
Now, MX (t) = (etb − eta )
t(b − a)
b2t 2 b3t 3 a2t 2 a3t 3
   
1
= 1 + bt + + + ... − 1 + at + + + ...
t(b − a) 2! 3! 2! 3!
(b2 − a2 )t 2 (b3 − a3 )t 3
 
1
= t(b − a) + + + ...
t(b − a) 2! 3!
b+a t b2 + ab + a2 t 2
= 1+ . + . + ...
2 1! 3 2!

13
 t a+b
Mean = µ1′ = coefficient of in MX (t) =
Γ1 2
t2 a2 + ab + b2
µ2′ = coefficient of in MX (t) =
Γ2 3

Now,Var(X) = µ2′ − (µ1′ )2

a2 + ab + b2 (a + b)2
= −
3 4
1
= (4a2 + 4ab + 4b2 − 3a2 − 3b2 − 6ab)
12
1
= (a − b)2
12

Problems
1.If X is uniformly distributed over (0,10) find (i)P(X < 4), (ii)P(X > 6), (iii)P(2 < X < 5).
Solution:
X is uniformly distributed over (0,10)

 1 , 0 < x < 10
∴ pdf is f (x) = 10
0, otherwise

Z 4
1 4 2
P(X < 4) = dx = =
0 10 10 5
Z 10
1 4 2
P(X > 6) = dx = =
6 10 10 5
Z 5
1 3
P(Z < X < 5) = dx =
2 10 10
1 1
2.Show that for the rectangular distribution dF = dx, o ≤ x < 1, µ1′ = and variance =
2 12
Solution:
dF
Given = 1, 0 ≤ x < 1,
dx
where F is cumulative function.

(
1, 0 ≤ x < 1
∴ F ′ (x) = 1 =⇒ f (x) =
0, otherwise
Z 1 Z 1
1
µ1′ = E(X) = x. f (x)dx = x dx =
0 0 2
1
x3
Z 1 
1
µ2′ = E(X ) = 2 2
x f (x)dx = =
0 3 0 3

2 1 1 1
∴ Var(X) = µ2′ − µ1′ = − =
3 4 12

14
3. Buses arrive at a specified stop at 15 minutes interval starting at 6 AM i.e. they arrive at 6 AM,
6.15 AM., 6.30 AM. and so on. If a passenger arrives at the stop at a time that is uniformly distributed
between 6 and 6.30 AM, find the probability that he waits (i) less than 5 minutes for a bus. (ii) more
than 10 minutes for a bus.
Solution:
Let X denote the number of minutes after 6 AM that a passenger arrives at the stop.
X is a uniform R.V over the interval (0,30)
1
∴ its pdf is f (x) = , 0 < x < 30
30
(i) Since the buses arrive at 15 minutes interval strating with 6 AM, a passenger has to wait less than
5 minutes if he comes to the stop between 6.10 and 6.15 or between 6.25 and 6.30 AM.

∴ Required probability = P(10 < X < 15) + P(25 < X < 30)
R 15 1 R 30 1
= 10 dx 25 dx
30 30
5 5 1
= + =
30 30 3
(ii) The passenger has to wait more than 10 minutes if he comes to the stop between 6 and 6.5 AM or
between 6.15 and 6.20 AM.

∴ Required probability = P(0 < X < 5) + P(15 < X230)

R5 1 R 20 1 1
= 0 dx 15 dx =
30 30 3
4. A random variable X has a uniform distribution over the interval (-3,3). Compute (1) P(X =
1
2) (2) P(|X − 2| < 2) (3) Find k such that P(X > k) = .
3
Solution:
Given X has a uniform distribution
 over the interval (-3,3).
 1 , −3 < x < 3
∴ the pdf of X is f (x) = 6
0, otherwise


Since X is continuous random variable, P(X = 2) = 0

P(|X − 2| < 2) = P(−2 < X − 2 < 2)

= P(−2 + 2 < X < 2 + 2) = P(0 < X < 4)

Z 4 Z 3
1
= f (x) dx = dx
0 0 6

1 3 3 1
= [x] = =
6 0 6 2

R∞ R3 1 1 3 3−k
(3) P(X > k) = k f (x)dx = k dx = [x] =
6 6 k 6
1
GivenP(X > k) =
3

15
 3−k 1
=⇒ =
6 3
=⇒ 3 − k = 2
∴k=1

5 Exponential Distribution
Definition:
A continuous RV X is said to follow an exponential distribution or negative exponential distribution
with prameter λ > 0, if its probability density function is given by
(
λ eλ x x ≥ 0
f (x) =
0 otherwise

λ eλ x dx = 1 and hence f (x) is a legitimate density function.

R∞ R∞
We note that 0 f (x)dx = 0

5.1 Moment generating function, Mean and variance

MX (t) = E etX = −∞
  R ∞ tx
e f (x)dx

R ∞ tx −λ x
= 0 e λe dx

R ∞ −(λ −t)x
= λ 0 e dx
" #−∞
e−(λ −t)x
= λ , if λ − t > 0 =⇒ λ > t
−(λ − t)
0
λ
= [0 − 1]
−(λ − t)
λ
MX (t) = ,λ >t
λ −t
1 t −1

t t2 t3
MX (t) = = 1 − = 1 + + + + ...
1 − λt λ λ λ2 λ3
t 1
Mean = µ1′ = coefficient of in MX (t) =
Γ1 λ
t 2 2
µ2′ = coefficient of in MX (t) = 2
Γ2 λ

Now,Var(X) = µ2′ − (µ1′ )2

2 1 1
= 2
− 2= 2
λ λ λ

16
5.2 Memoryless Property of the Exponential Distribution
If X is exponentially distributed, then

P(X > s + t/X > s) = P(X > t), for anys,t > 0
Z ∞
P(X > k) = λ e−λ x dx
k

= (−eλ x )∞
k

= e−λ k (1)

P {X > s + t and X > s}

Now P(X > s + t/X > s) =
P {X > s}
P {X > s + t}
=
P {X > s}

e−λ (s+t)
= , [by (1)]
e−λ s

= e−λt = P(X > t).

Problems
1 −x
1.Find the moment generating function of the exponential distribution f (x) = e c , x ≥ 0 and c > 0.
c
Hence find its mean and S.D.
Solution:
We know for exponential distribution

f (x) = λ e−λ x , x ≥ 0, λ > 0

1
MX (t) = t
1−
λ
1 1
Mean = Var(X) = 2
λ λ
1
In the given problem λ =
c
1
∴ MX (t) =
1 − ct

∴ Mean = c

Var(X) = c2

∴ S.D. = c

17
2.Suppose the duration X in minutes of long distance calls from your home follows exponential law
with pdf 
 1 e−x/5 for x > 0
f (x) = 5
0 otherwise


find P(X > 5), P(3 ≤ X ≤ 6), mean and variance of X.

Solution:
We know that for an exponential distribution given by
(
λ e−λ x , x > 0
f (x) =
0 otherwise

1
mean = , Var(X)=25
λ

1
Z ∞ Z ∞
P(X > 5) = f (x)dx = e−x/5 dx
5 5 5
 ∞
1  e−x/5 
=
5 −1
 
5 5

= −(0 − e−1 )
1
= = 0.36788
e

Z 6
P(3 ≤ X ≤ 6) = f (x)dx
3
Z 6
1
= e−x/5 dx
5 3
 6
1  e−x/5 
=
5 −1
 

h −65 −3 3
i
= − e −e
5 5

−3 −6
= e 5 −e 5
= 0.5488 − 0.3012 = 0.2476

3.Suppose the length of life of an appliance has an exponential distribution with mean 10 years.
A used appliance is bought by a person. What is the probability that it will not fail in the next 5 years?
Solution:
Let X represent the length of time of the appliance

18
Given X follows exponential distribution with mean =10
∴ its pdf f (x) = λ e−λ x , x ≥ 0
1
Mean = = 10
λ
1
=⇒ λ =
10
1 −x
∴ f (x) = e 10 , x ≥ 0
10
Since exponential distribution has memoryless property, the length of life t years before the purchase
is irrelevant.
∴ P(X > t + 5|X > t) = P(X > 5)
Z ∞
= f (x)dx
5
1 ∞ −x
Z
= e 10 dx
10 5 ∞
−x
1  e 10 
=
10 −1
 
10 5
−1
= e 2 = 0.6065

1
4.The time required to repair a machine is exponentially distributed with parameter . What is the
2
conditional probability that the repair exceeds 2 hours? What is the conditional probability that the
repair takes atleast 10 hours given that the duration exceeds 9 hours?
Solution:
Let X denote the repairing time.
1
Given X follows exponential distribution with parameter λ =
2
∴ The exponential distribution is given by
f (x) = λ e−λ x , x ≥ 0

1 −1 x
= e 2 ,x≥0
2
Z ∞
P(X > 2) = f (x)dx
2

1
Z ∞
= e−x/2 dx
2 2
" #
1 e−x/2
=
2 −1/2

= − e−∞ − e−1
 

1
= e−1 =
e
P(X ≥ 10 ∩ X > 9)
RequiredP(X ≥ 10|X > 9) = = P(X > 1)
P(X > 9)

19
 R∞
= 1 f (x)dx

1 −1x
Z ∞
= e 2 dx
1 2
 ∞
1
1 e− 2 x
=

 −1 
2
2 1
 
−∞ −1/2
= − e −e

= e−0.5 = 0.6

6 Normal distribution
Definition
A continuous RV X is said to follow normal distribution or Gaussian distribution with parameter µ
and σ , if its pdf is given by
1 2 2
f (x) = √ e−(x−µ) /2σ ; −∞ < x < ∞, σ > 0
σ 2π

Note:
t2σ 2
1. Moment generating function of Normal distribution is MX (t) = eµt+ 2 .

2. Mean = µ.

3. Variance, Var(X) = σ 2 .

Properties of Normal distribution

1. The mean, median and mode coincide at x = µ.

2. It is symmetric about x = µ.
1
3. The maximum value of f (x) is at x = µ and the maximum value = √ .
σ 2π
4. The points of inflection are at x = µ − σ and x = µ + σ .

5. The curve approaches the horizontal axis asymptotically on either side of x = µ.

6. The total area under the curve is 1.

7. Coefficient of skewness is 0 and coefficient of kurtosis is 3.

8. All odd moments about mean vanish.

20
Standard Normal Distribution

X −µ
Z=
σ
1 1
E(Z) = (E(X) − µ) = (µ − µ) = 0
σ σ
1 1 2
Var(Z) = 2 Var(X) = 2 σ = 1
σ σ
1 −z2
∴ Its pdf is given by φ (z) = √ e 2 , −∞ < z < ∞.
2π
1 2
Its MGF is given by MZ (t) = e 2 t .

Problems
1.X is a normal distribution with mean 16 and S.D. 3. Find (i)P(X ≥ 19), (ii)P(12.5 < X < 19), (iii)P(10 <
X < 25), (iv)K if P(X > K) = 0.24
Solution:
Given µ = 16, σ = 3
X −µ X − 16
Put Z = =
σ 3
19 − 16
(i) when X = 19, Z = =1
3

∴ P(X ≥ 19) = P(Z ≥ 1) = 0.5 − P(0 < Z < 1) = 0.5 − 0.3413 = 0.1587

(ii) P(12.5 < X < 19)

12.5 − 16
when X = 12.5, Z = = −1.16
3
19 − 16
when X=19, Z= =1
3

∴ P(12.5 < X < 19) = P(−1.16 < Z < 1)

= P(−1.16 < Z < 0) + P(0 < Z < 1)

= P(0 < Z < 1.16) + P(0 < Z < 1), (by symmetry)

= 0.3770 + 0.3413 = 0.7183

(iii) P(10 < X < 25)

10 − 16
when X=10, Z= = −2
3

21
 25 − 16
when X=25, Z= =3
3
∴ P(10 < X < 25) = P(−2 < Z < 3)

= P(−2 < Z < 0) + P(0 < Z < 3)

= P(0 < Z < 2) + P(0 < Z < 3), (by symmetry)

= 0.4772 + 0.4987 = 0.9759

(iv)P(X > K) = 0.24

K−µ K − 16
when X=K, Z= = = z1 (say)
σ 3
∴ P(X > K) = 0.24

=⇒ P(Z > z1 ) = 0.24

=⇒ P(0 < Z < z1 ) = 0.5 − 0.24 = 0.26

∴ z1 is the value of Z corresponding to the area 0.26.

and z1 = 0.7(from table)
K − 16
∴ = 0.7
3
=⇒ K = 16 + 3 × 0.7 = 16 + 2.1 = 18.1
2. The average percentage of marks of candidates in an university examination is 42 with S.D of
10. The maximum for a pass is 50%. If 1000 candidates appear for the examination, how many can
be expected to pass it assuming normality of the distribution of marks? If it is required that double
that number should pass, what should be the minimum percentage of marks?
Solution:
Given µ = 42, σ = 10
Let X ∼ N(µ, σ 2 )

X −µ X − 42
Z= =
σ 10
A student is passed when scored greater than 50 i.e. X ≥ 50
50 − 42
∴ when X = 50, then Z = = 0.8
10
to ∴ P(X ≥ 50) = P(Z ≥ 0.8)
=0.5-P(0¡Z¡0.8)
=0.5-0.2881
=0.2119

Number of students expected to pass =1000 × 0.2119=211.9 ≈212

424
If the no. of passed student is 424, then proportion out of 1000= = 0.424
1000

22
Let X = x, is minimum pass marks and corresponding Z = z1 , then

P(Z > z1 ) = 0.424

0.5 − P(0 < Z < z1 ) = 0.424
∴ P(0 < Z < z1 ) = 0.5 − 0.424 = 0.076
Now from the table, z1 = 0.19
X − 42
= 0.19
10
X = 0.19 × 10 + 42 = 43.9 ≈ 44

∴ The minimum pass marks to double pass student is 44%.

3. In a distribution exactly normal 7% of the items are under 35 and 89% are under 63. What
are the mean and S.D. of the distribution?(Draw diagram if required)
Solution:
Let X ∼ N(µ, σ 2 )

P(X < 35) = 7% = 0.07

P(X < 63) = 89% = 0.89
X −µ
Let Z =
σ
35 − µ
When X = 35, X = = −Z1
σ
35 − µ = −σ Z1 (1)
Now,

P(X < 35) = P(Z < −Z1 ) = 0.07

P(Z > Z1 ) = 0.07
0.5 − P(0 < Z < Z1 ) = 0.07
∴ P(0 < Z < Z1 ) = 0.43
∴ Z1 = 1.48 From table
63 − µ
When X = 63, Z = = Z2 , since area below X = 63 is 0.69.
σ
The value of Z is positive
63 − µ = σ Z2 (2)

P(X < 63) = 0.89

P(Z < Z2 ) = 0.89 = 0.5 + 0.39
∴ P(0 < Z < Z2 ) = 0.39

Z2 is the value of Z corresponding to the area =0.39

From table Z2 = 1.23

From (1), 35 − µ = 1.48σ (3)
From (2), 63 − µ = 1.23σ (4)

23
Solving (3) and (4), we have

µ = 50.3, σ = 10.33

4. In a normal distribution, 31% of the items are under 45 and 8% are over 64. Find the mean
and variance of the distribution.(Draw diagram if required)
Solution:
X −µ
We know that, Z =
σ
Let Z = Z1 , where X = 45 and Z = Z2 , when X = 64

P(0 < Z < Z1 ) = 0.31

P(Z1 < Z < 0) = 0.19
∴ From tables, Z1 = −0.49
45 − µ
i.e., = −0.49 or
σ
45 − µ = −0.49σ (1)
P(Z > Z1 ) = 0.8 or
P(0 < Z < Z2 ) = 0.42
∴ From tables, Z2 = 1.40
64 − µ
∴ = 1.40
σ
64 − µ= 1.40σ (2)
(1) − (2) =⇒ −19 = −1.89σ
−19
σ = = 10
−1.89
From (1) µ = 45 + (0.49 × 10) = 45 + 4.9 = 49.9 = 50

Mean = 50 and S.D = 10

i.e.,µ = 50, σ = 10, Variance(σ 2 ) = 100

24