Probability distributions (Module - 2)

BAMAT207

S. Devi Yamini

Module - II 1 / 63
Overview

1 Discrete Distributions
Bernoulli Distribution
Binomial Distribution
Poisson Distribution
Geometric Distribution
Negative Binomial Distribution

Module - II 2 / 63
Overview

1 Discrete Distributions
Bernoulli Distribution
Binomial Distribution
Poisson Distribution
Geometric Distribution
Negative Binomial Distribution
2 Continuous Distributions
Exponential Distribution
Normal Distribution
Gamma Distribution
Weibull Distribution
Uniform Distribution

Module - II 2 / 63
Overview

1 Discrete Distributions
Bernoulli Distribution
Binomial Distribution
Poisson Distribution
Geometric Distribution
Negative Binomial Distribution
2 Continuous Distributions
Exponential Distribution
Normal Distribution
Gamma Distribution
Weibull Distribution
Uniform Distribution
3 Sampling Distributions

Module - II 2 / 63
Discrete Distributions

Bernoulli Distribution
Binomial Distribution
Poisson Distribution
Geometric Distribution
Negative Binomial Distribution

Module - II 3 / 63
Bernoulli Distribution

A Bernoulli trial is a random experiment in which there are exactly two

possible outcomes - success or failure.

Example
Head or Tail ; Defective or Non-defective; Suvival or Death; True or False;
etc.

Let X denote the random variable representing success. Then X = {0, 1}

P(X = 1) = p P(X = 0) = 1 − p = q
X
Mean E (X ) = xP(x) = 1 × p + 0 × q = p

Variance V (X ) = E (X 2 ) − (E (X ))2 = p(1 − p) = pq

Module - II 4 / 63
Problems on Bernoulli Distribution

Fit a Bernoulli dsitribution for the following data:

Outcome (x) 0 1
Frequency (f ) 18 42

Solution:
Total number of observations: 18 + 42 = 60
Probability of success = p = 4260 = 0.7
P(X = 1) = 0.7, P(X = 0) = 1 − p = 0.3
Hence, the fitted distribution is
(
0.3 x =0
P(X = x) =
0.7 x =1

X ∼ Bernoulli(0.7)

Module - II 5 / 63
Binomial Distribution

Binomial distribution results when it satisfies the following:

- It has n repeated Bernoulli trials (i.e. outcomes are either success or
failure in each trial)
- The trials are independent
- The probability of success in each trial is constant

Notations
n Number of trials
X Number of successes in n trials
p Probability of success in each trial
q Probability of failure in each trail
P(X = x) Probability of exactly x scucesses among n trails
P(X ≤ x) Probability of at most x successes
P(X ≥ x) Probability of at least x successes

Module - II 6 / 63
Binomial Distribution

Since X denotes the number of successes in n trials, X = {0, 1, 2, . . . , n}.

The probability mass function of the binomial distribution is

P(X = x) = nCx p x q n−x x = 0, 1, 2, . . . , n

X ∼ Binomial(n, p)

Mean and Variance of a binomial distribution

Mean E (X ) = np

Variance V (X ) = npq

Module - II 7 / 63
Problems
1. A biased coin is tossed 6 times. The probability of heads on any toss is
0.3. Let X denote the number of heads that come up. Calculate (i)
P(X = 2), (ii) P(X ≤ 3), (iii) P(1 < X ≤ 5).

Module - II 8 / 63
Problems
2. The incidence of a certain disease is such that 20% of workers suffer
from it. If 10 workers are selected at random, find the probability that (i)
exactly 2 workers suffer from the disease, (ii) not more than 2 workers
suffer from the disease. ( (i) 0.302, (ii) 0.678 )
3.
√ For a binomial distribution, the mean is 6 and the standard deviation is
2. Find P(X = 0). (0.00005)
4. The probability of a man hitting a target is 14 . How many times must
he fire so that the probability of hitting the target at least once is more
than 32 ? (n = 4 approx. )
5. Bottled sweet milk stored in a godown is reported to have gone sour. A
test check has revealed that milk in 25% of the bottles is bad and thus
unfit for consumption. The salesman at a retail outlet offers 5 bottles for
sale on demand. Find the probability that milk will be unfit for
consumption (a) exactly in 2 bottles, (b) at least in 2 bottles, (c) at most
in 2 bottles, and (d) between 2 and 4 bottles. ((a) 0.2636, (b) 0.3672, (c)
0.8964, (d) 0.3662)
Module - II 9 / 63
Solutions
2. X denote the number of workers suffering from the disease.
n = 10, p = 0.2, q = 0.8

X ∼ Binomial(10, 0.2)

(i)P(X = 2) = 10C2 (0.2)2 (0.8)8 = 0.302

(ii)P(X ≤ 2) = P(X = 0) + P(X = 1) + P(X = 2) = 0.678
3. √
√
Mean = np = 6 SD = npq = 2
So,
1 2
npq = 2 =⇒ q = , p=
3 3
Also, n = 9
2 1 1
P(X = 0) = 9C0 ( )0 ( )9 = 9
3 3 3

Module - II 10 / 63
Solutions

4. Probability of hitting the target in one shot: p = 14

3
Probability of missing the target in one shot: q = 1 − p = 4
Let the man fire n times.
3
P(at least one hit) = 1 − P(no hit) = 1 − ( )n
4
3 2
1 − ( )n >
4 3
3 1
( )n <
4 3
Taking logarithm on both sides, we obtain n = 4 approximately.

(Try Problem 5 !)

Module - II 11 / 63
6. Fit a binomial distribution to the following data.

Number of successes (x) 0 1 2 3

Frequency (f ) 10 25 30 15

Solution:
Note that n = 3 and Total frequencyPN = 80
The mean of the above table is x̄ = Nxf = 1.625 = np (mean of binomial
distribution)
Hence, p = 1.625
3 = 0.542, q = 0.458

x P(X = x) = nCx p x q n−x Expected frequency

0 3C0 (0.542)0 (0.458)3 = 0.096 80 × 0.096 = 7.68 ∼ 8
1 3C1 (0.542)1 (0.458)2 = 0.342 80 × 0.342 = 27.36 ∼ 27
2 3C2 (0.542)2 (0.458)1 = 0.404 80 × 0.404 = 32.32 ∼ 32
3 3C3 (0.542)3 (0.458)0 = 0.158 80 × 0.158 = 12.64 ∼ 13

Module - II 12 / 63
Poisson Distribution

Module - II 13 / 63
Probability mass function

Mean = Variance= λ

Module - II 14 / 63
Problems

1. Consider a computer system with Poisson job arrival stream at an

average of 2 per minute. Determine the probability that in any one minute
interval there will be (i) zero jobs, (ii) exactly 3 jobs, (iii) at most 3 jobs,
(iv) less than 3 job arrivals. ( (i) 0.135, (ii) 0.18, (iii) 0.8571, (iv) 0.6767)
2. The number of calls coming per minute into a hotel reservation center
is a poisson random variable with mean 3. Find the probability that (i) no
calls come in a given one minute period, (ii) at least two calls will arrive in
a given two minute period. ( (i) 0.04979, (ii) 0.9826)
3. In a certain poisson distribution, the frequency corresponding to two
successes is half the frequency corresponding to three successes. Find the
mean and standard deviation. ( Mean=6, SD=2.45)
4. Calculate P(X = 3) given that the variance of X is 0.5. (0.0126)

Module - II 15 / 63
Solutions

1. Let X denote the number of jobs arriving in one minute.

λ = 2 per minute

e −2 2x
P(X = x) =
x!
(i)P(X = 0) = e −2 = 0.135
(ii)P(X = 3) = 0.18
(iii)P(X ≤ 3) = P(X = 0) + P(X = 1) + P(X = 2) + P(X = 3) = 0.8571
(iv )P(X < 3) = P(X = 0) + P(X = 1) + P(X = 2) = 0.6767

Module - II 16 / 63
Solutions

2. Let X denote the number of calls in one minute.

λ = 3 per minute
e −3 3x
P(X = x) =
x!
(i)P(X = 0) = e −3 = 0.0498
(ii) In two minutes, the mean λ = 3 × 2 = 6 per 2 minutes.

P(X ≥ 2) = 1 − P(X < 2) = 1 − (P(X = 0) + P(X = 1))

= 1 − e −6 (1 + 6) = 0.9826

Module - II 17 / 63
Solutions

3.
1
P(X = 2) = P(X = 3)
2
e −λ λ2 1 e −λ λ3
=
2! 2 3!
Simplification leads to λ = 6 =Mean=Variance
√
Standard deviation= 6.
4. λ = 0.5
e −0.5 0.53
P(X = 3) = = 0.0126
3!

Module - II 18 / 63
Problems
5. Fit a Poisson distribution to the following data giving the number of
defects per item.
Number of defects (x) 0 1 2 3 4
Frequency (f ) 40 35 15 8 2

Solution:
TotalPfrequency N = 100
x̄ = Nxf = 0.97 = λ (mean of the Poisson distribution)
e −λ λx
x P(X = x) = x! Expected frequency
0 0.379 100 × 0.379 = 37.9 ∼ 38
1 0.368 100 × 0.368 = 36.8 ∼ 37
2 0.178 100 × 0.178 = 17.8 ∼ 18
3 0.058 100 × 0.058 = 5.8 ∼ 6
4 0.014 100 × 0.014 = 1.4 ∼ 1

Module - II 19 / 63
Geometric Distribution

In a series of Bernoulli trials (independent trials with constant probability

p of a success), the random variable X that equals the number of trials
until the first success is a geometric random variable with parameter p and

P(X = x) = (1 − p)x−1 p x = 1, 2, . . .

A geometric random variable has been defined as the number of trials until
the first success.
1 (1−p)
Mean = p Variance = p2

P(X > x) = (1 − p)x

Module - II 20 / 63
Memoryless property

Geometric distribution is the only discrete distribution that has memoryless

property.

If X (number of trials until the first success) follows geometric distribution

with parameter p, then for all integers m, n ≥ 1,

P(X > m + n|X > m) = P(X > n)

Proof

P(X > m + n) ∩ P(X > m) P(X > m + n)

P(X > m + n|X > m) = =
P(X > m) P(X > m)

(1 − p)m+n
P(X > m + n|X > m) = = (1 − p)n = P(X > n)
(1 − p)m

Module - II 21 / 63
Problems
1. The probability that a wafer contains a large particle of contamination
is 0.01. If it is assumed that the wafers are independent, what is the
probability that exactly 125 wafers need to analyzed before a large particle
is detected?
Solution: Let X denote the number of samples analyzed until a large
particle is detected. X follows a geometric random variable with p = 0.01
P(X = 125) = (0.99)124 × 0.01 = 0.0029

2. If the probability is 0.05 that a certain kind of measuring device will

show excessive drift, what is the probability that the sixth measuring
device tested will be the first to show excessive drift?
Solution: Let X denote the number of measuring devices tested until the
first one shows excessive drift.
p = 0.05
P(X = 6) = (0.95)5 × (0.05) = 0.0387

Module - II 22 / 63
A problem based on memoryless property

3. A machine produces items, each of which is defective with probability

0.1, independently of the others. Items are inspected one by one. Let X
denote the number of items inspected until the first defective item is
found. Find P(X > 10|X > 5).

Solution
By memoryless property of geometric distribution,

P(X > 10|P(X > 5) = P(X > 5) = (1 − p)5 = (1 − 0.1)5 = 0.5904

Module - II 23 / 63
4. Fit a geometric distribution to the following data showing the number
of trials until the first success.
Number of trials (x) 1 2 3 4 5 ≥6
Frequency (f ) 20 30 25 15 10 0

Solution
The total frequency
P N = 100
xf
Mean x̄ = N = 2.65
1
Mean of a geometric distribution = p = 2.65
p = 0.377

x P(X = x) = (1 − p)x−1 × p Expected frequency

1 (1 − 0.377)0 × 0.377 = 0.377 100 × 0.377 = 37.7 ∼ 38
2 (1 − 0.377)1 × 0.377 = 0.235 100 × 0.235 = 23.5 ∼ 23
3 (1 − 0.377)2 × 0.377 = 0.147 100 × 0.147 = 14.7 ∼ 15
4 (1 − 0.377)3 × 0.377 = 0.092 100 × 0.092 = 9.2 ∼ 9
5 (1 − 0.377)4 × 0.377 = 0.057 100 × 0.057 = 5.7 ∼ 6
1 − 5x=1 P(X = x) = 0.092
P
≥6 100 × 0.092 = 9.2 ∼ 9
Module - II 24 / 63
Problems

4. An automated weight monitor can detect underfilled cans of beverages

with probability 0.98. What is the probability it fails to detect an
underfilled can for the first time when it encounters the 10th underfilled
can? (Here X denotes the number of underfilled cans encountered until
the first failure of detection. Hence,
P(X = 10) = (1 − 0.02)9 × 0.02 = 0.0167)

5. A company manufactures hydraulic gears, and records show that the

probability is 0.04 that one of its new gears will fail its inspection test.
What is the probability that the fifth gear in a day will be the first one to
fail the test? (Here, X denotes the number of gears tested until the first
failure. Hence, P(X = 5) = (1 − 0.04)4 × 0.04 = 0.034)

Module - II 25 / 63
Negative Binomial Distribution

Negative binomial distribution is the distribution of the number of

trials needed to get the rth success (fixed number of successes).
For the r th success to occur in the xth trial, the first x − 1 trials must
result in r − 1 successes.

Module - II 26 / 63
Negative Binomial Distribution

Negative binomial distribution is the distribution of the number of

trials needed to get the rth success (fixed number of successes).
For the r th success to occur in the xth trial, the first x − 1 trials must
result in r − 1 successes.

Let X denote the random variable representing the trial at which the r th
success occurs, where r is a fixed integer. The probability of r th success in
xth trial is

P(X = x) = (x − 1)Cr −1 p r (1 − p)x−r where x = r , r + 1, . . .

Module - II 26 / 63
Negative Binomial Distribution

Negative binomial distribution is the distribution of the number of

trials needed to get the rth success (fixed number of successes).
For the r th success to occur in the xth trial, the first x − 1 trials must
result in r − 1 successes.

Let X denote the random variable representing the trial at which the r th
success occurs, where r is a fixed integer. The probability of r th success in
xth trial is

P(X = x) = (x − 1)Cr −1 p r (1 − p)x−r where x = r , r + 1, . . .

r
Mean =
p
r (1 − p)
Variance =
p2
Module - II 26 / 63
Negative Binomial Distribution

Problem
A person conducting telephone surveys must get 3 more completed surveys
before their job is finished. On each random dialed number, there is a 9%
chance of reaching an adult who will complete the survey. What is the
probability that the 3rd completed survey occurs on the 10th call?

Module - II 27 / 63
Negative Binomial Distribution

Problem
A person conducting telephone surveys must get 3 more completed surveys
before their job is finished. On each random dialed number, there is a 9%
chance of reaching an adult who will complete the survey. What is the
probability that the 3rd completed survey occurs on the 10th call?

Solution
p = 0.09 is the probability that a call results in a completed survey.
For the 3rd success to occur on the 10th call, there must be exactly 2
successes in the first 9 calls and the 10th call must be a success.
Also, r = 3, x = 10.

P(X = 10) = 9C2 (0.09)3 (0.91)7 = 0.015

So, there is about 1.5% chance that 3rd completed survey happens exactly
on the 10th call.
Module - II 27 / 63
Continuous Distributions

Exponential distribution

Normal distribution

Uniform distribution

Gamma distribution

Weibull distribution

Module - II 28 / 63
Exponential Distribution

Module - II 29 / 63
Exponential Distribution

1
Mean =
θ
1
Variance =
θ2

Graph of exponential distribution

Module - II 30 / 63
Exponential Distribution

Module - II 31 / 63
Exponential Distribution

This property shows that past failures do not affect the future probabilities.
Example
If a student has already failed 5 times, and the probability of passing
within the next 3 attempts is the same as probability of passing within the
first 3 attempts. The process ”starts afresh” after each attempt. This
clearly illustrates the memoryless property of geometric distribution.

P(X ≥ 8|X ≥ 5) = P(X ≥ 3)

Module - II 32 / 63
Problems

Module - II 33 / 63
Solution

Module - II 34 / 63
Solution

Module - II 35 / 63
Solution

Problem 4
Students arrive at a local restaurant according to an approximate Poisson
process at a mean rate of 30 students per hour. What is the probability
that the waiter has to wait for more than 3 mins for the next student?

Solution
Arrivals follow a Poisson distribution with mean λ = 3060 per min.
Let X denote the waiting time for the next arrival which follows an
exponential distribution.
So, the mean of the exponential distribution is 1θ = λ1 = 2
Hence, θ = 12

P(X > 3) = e −3θ = e −1.5

Module - II 36 / 63
Normal Distribution

Module - II 37 / 63
Normal Distribution

Module - II 38 / 63
Standard distribution table

Module - II 39 / 63
Problems

Module - II 40 / 63
Problems

Module - II 41 / 63
Problems

Module - II 42 / 63
Problems

Module - II 43 / 63
Gamma Distribution

Module - II 44 / 63
Gamma Distribution

Module - II 44 / 63
Gamma Distribution

Properties
1 Gamma distribution is a generalization of exponential distribution.
When α = 1, we obtain the pdf of exponential distribution.

2 Scale parameter- λ; Shape parameter- α

3 Mean = αλ ; Variance = α
λ2

1
4 β= λ

Module - II 45 / 63
Gamma Distribution

Module - II 46 / 63
Problems

Module - II 47 / 63
Problems

Module - II 48 / 63
Problems

Module - II 49 / 63
Weibull Distribution

Module - II 50 / 63
Weibull Distribution

β
P(X ≤ a) = 1 − e −αx
β
P(X > a) = e −αx

Module - II 51 / 63
Weibull Distribution

Module - II 52 / 63
Try these!
1 The lifetime X , measured in hours, of a spring that is operated
continuously has the cumulative distribution function
( k
1 − e −λx , x > 0,
F (x) =
0, otherwise.
where λ = 0.001 and k = 2. Find (a) the probability that a new
spring survives 30 hours and (b) the probability that the spring that
has been operating without failure for 50 hours survives an additional
30 hours.
[Hint: (a) P(X > 30), (b) P(X > 80|X > 50)]
2 A magnetic disk is exposed to corrosive gas. Lifetime of a disk follows
Weibull distribution with parameters α = 300 and β = 0.5. What will
be the probability of it (i) failing before 500 hours? (ii) lasts 600
hours or more?
3 The length of life (in hours) of a drill bit in a mechanical operation
has a Weibull distribution with α = 2 and β = 50. Find the
probability that the bit will fail before 10 hours of usage.
Module - II 53 / 63
Uniform Distribution

The Uniform or Rectangular distribution has a random variable X

restricted to a finite interval [a, b] and has
(
1
, a ≤ x ≤ b,
f (x) = b−a
0, otherwise.

Module - II 54 / 63
Mean and Variance

b+a
E (X ) =
2
(b − a)2
V (X ) =
12

Module - II 55 / 63
Problem

A bus is uniformly late between 2 and 10 minutes. How long can you
expect to wait? If its more than 7 minutes late, you will be late for work.
What is the probability of you being late?
Solution:
a = 2, b = 10
X ∼ U(2, 10)
a+b
Expected waiting time = E (X ) = 2 = 6. To find P(X > 7).
Z 10
1
P(X > 7) = dx
7 b−a
3
P(X > 7) = = 0.375
8

Module - II 56 / 63
Problems

1. The continuous random variable X is uniformly distributed over the

interval [-4,6]. Find (a) the mean of X , (b) P(X ≤ 2.4), (c)
P(−3 < X − 5 < 3).

2. Packages have a nominal net weight of 1 kg. However their actual net
weights have a uniform distribution over the interval 980 g to 1030 g.
(a) Find the probability that the net weight of a package is less than 1 kg.
(b) Find the probability that the net weight of a package is less than w g,
where 980 < w < 1030.

3. If X is a random variable having a uniform distribution U(a, b) such

that P(20 < X < 40) = 0.2 and mean= 150, find the values of a and b.

Module - II 57 / 63
Sampling Distributions (Concepts only)

Supoose we take a sample of n = 10 observations from a population

and compute the sample average, say x̄ = 10.2.
Now we repeat this process, taking a second sample of n = 10
observations from the same population and the resulting sample mean
is 10.4.
The sample mean depends on the observations in the sample which
differs from sample to sample. The sample mean is a random variable
The probability distribution of a statistic (can be sample mean or
sample standard deviation or sample proportion or sample size) is
called the Sampling distribution
If X1 , X2 , . . . , Xn are independent and identically distributed random
variables with mean µ and variance σ 2 , then X̄ has mean E (X̄ ) = µ
2
and Var (X̄ ) = σn .

Module - II 58 / 63
Central limit theorem

If X1 , X2 , . . . , Xn is a random sample of size n taken from a population

with mean µ and variance σ 2 , and if X̄ is the sample mean, then for large
2
n, the sampling distribution of X̄ ∼ N(µ, σn ).
That is,
X̄ − µ
Z= σ
√
n

is the standard normal distribution as n 7→ ∞.

Module - II 59 / 63
Sampling Distributions

Students’ t-distribution
F- distribution
Chi-square distribution

Module - II 60 / 63
Students’ t-distribution

Module - II 61 / 63
χ2 distribution

If Z1 , Z2 , . . . , Zk are independent standard normal random variables,

then
Z12 + Z22 + . . . , +Zk2
has a χ2 distribution with k degrees of freedom.
The pdf of χ2 distribution with k degrees of freedom is
k x
x 2 −1 e − 2
f (x) = k
2 2 Γ( k2 )

for x ≥ 0

Module - II 62 / 63
F-distribution

A continuous distribution with the following pdf where ν1 and ν2 are

degrees of freedom:
ν1 ν1
−1
Γ( ν1 +ν ν1 2
2 )( ν2 ) x
2 2
f (x) = ν1 +ν2
Γ( ν21 )Γ( ν22 )(1 + ν1
ν2 x)
2

for x ≥ 0.

Module - II 63 / 63