PROBABILITY

DISTRIBUTIONS
 Course Objective
To understand
 Random Variable

 Frequency Distributions

 Probability Distributions
 Binomial Distribution
 Poisson Distribution
 Normal Distribution
 Random Variable (RV)
 A variable that can take many real values which are
determined by the outcomes of a random
experiment on a real line -∞,+∞
 A chance or Stochastic Variable
 Value depends on outcomes of a random experiment
 Discrete (countable finite values) or Continuous
(uncountable or infinite values)
 Relative Frequency of a
Random Variable
Consider the following frequency distribution

No. of
items 0 1 2 3 4 5 Total
sold X :

Number
of days= 10 15 30 20 15 10 100
f(x)
 Relative Frequency of a
Random Variable
Now prepare a table of relative frequency
(divide each frequency by the sum of frequencies)

Number of items sold Relative Frequency

X
0 10/100
1 15/100
2 30/100
3 20/100
4 15/100
5 10/100
 Probability distribution table
 Values from previous step yield the following
results
 The below table is the probability distribution table
obtained from the relative frequencies
X=xi 0 1 2 3 4 5

P(X=xi) 0.1 0.15 0.3 0.2 0.15 0.1

 Mean, Variance and Standard
Deviation of RV
 Mean or Expected Value of a RV:
μ = E(X) = {x1p1+x2p2+……xnpn} = ∑pixi
(For i = 1 to n)
 Variance of a RV:
σ2 = V(X) = E(Xi2)- [E(Xi)]2
where E(Xi2) = x12p1+ x22p2+…… xn2pn =∑xi2pi (For i = 1 to n)
 Standard deviation of a RV:
σ = S.D= +√ V(X) = √ E(Xi2)- [E(Xi)]2
 Problem 1
 A dealer in Air coolers estimates from his past
experience the probabilities of selling Air coolers in
a day as given below. Find Mean and S.D

No.
0 1 2 3 4 5 6
(X)

Pblty
0.02 0.10 0.21 0.32 0.20 0.09 0.06
(P)
 Solution
Mean= μ = E(X) = ∑pixi
= x1p1+x2p2+ x3p3+x4p4+ x5p5+x6p6

= 0 X 0.02 + 1 X 0.10 + 2 X 0.21 +

3 X 0.32 + 4 X 0.20 + 5 X 0.09 +
6 X 0.06

= 0+ 0.1+0.42+0.96+0.80+0.45+0.36

μ = 3.09
 Problem 1 - Solution
Variance = σ2 = V(X) = E(Xi2)- [E(Xi)]2

=
{ 02(0.02) + 12(0.10) + 22(0.21) + 32(0.32) +
42(0.20) + 52(0.09) + 62(0.06)}- [3.09]2

={ 0 + 1(0.1) + 4(0.21) + 9(0.32) +

16(0.20) + 25(0.09)+ 36(0.06)} - [3.09]2

= { 0.1+0.84+2.88+3.2 +2.25 + 2.16 - [9.5481]

= { 11.43} - [9.5481] = 1.8819

_____
σ =√1.8819=1.3718
 Problem 2
A man runs an “ice cream” parlor in a holiday
resort. If the summer is “mild” he can sell 2500
cups of ice cream. If it is “hot” he can sell 4000
cups. If it is “ very hot” he can sell 5000 cups. It is
known that the probability in any year for the
summer to be mild is 1/7and hot is 4/7. A cup of
ice cream costs Rs. 2 and sold for Rs. 3.50. What is
his expected profit?
 Problem 2 - Solution
 Since sum of the probabilities equal to 1
{1/7 + 4/7 + P(Very hot) = 1
 P(Very hot) = 1 - {1/7 + 4/7} = 1 - 5/7 = 2/7
 Profit percup = Rs3.50-Rs2=Rs1.50

 Profit in a mild weather condition = 2500 X Rs.1.50 = Rs.3750

 Profit in hot weather condition = 4000 X Rs.1.50 = Rs.6000
 Profit in Very hot weather condition = 5000 X Rs.1.50 = Rs.7500

 Expected Profit = μ = E(X) = ∑pixi =

3750X1/7 + 6000 X 4/7 + 7500 X 2/7
= 6107.13
 Frequency Distributions
OBSERVED FREQUENCY THEORETICAL FREQUENCY
DISTRIBUTIONS DISTRIBUTIONS
 Distributions are based on  Distributions are based on
experimentation and actual mathematical models
observations  Derived mathematically
 Actual data is collected from some definite
and classified in the form probability law
of a frequency distribution
 Probability Distribution

 Probability distribution is the listing of all possible

outcomes of a random experiment together with
their probabilities
 Outcome of random experiment is a random
variable which may ne discrete or continuous
 Accordingly two types of probability distributions
are possible
 Types of Probability Distributions
DISCRETE PROBABILITY CONTINUOUS PROBABILITY
DISTRIBUTION DISTRIBUTION FUNCTION
 Probability distribution  A probability distribution
function defined on the function defined on the
discrete random variable (that continuous random
takes finite set of values, such
variable (that takes infinite
as number of defectives in
sample, no of printing values, such as income,
mistakes per page, number of height, age, weight,
accidents in a day, etc) thickness, length etc)
 Binomial probability  Normal probability
distribution and Poisson distribution
probability distribution
BINOMIAL DISTRIBUTION
(BD)
 BD
 Binomial Distribution B.D is also known as
Bernoulli distribution after the Swiss mathematician
James Bernoulli who discovered it in the year
1700A.D
 Deals with probabilities of trials having only two
possible outcomes (Bernoulli trials)
 S = { Success, Failure}; outcome of our interest is
Success and the other one is denoted as Failure
 BD - examples
 Tossing a coin has only two outcomes – Head
or Tail. If our interest is Head then getting a
Head is Success and getting a Tail is Failure
 Quality inspection yields two types of output –
Defective or Non-defective
 In a class room a student may be Present or
Absent
 BD - Assumptions
 Probability of Success is ‘p’ ; probability of
Failure is ‘q’ or ‘1 – p’ as there are only two
possible outcomes
 Each trial is independent of other trials. Hence,
p & q remain constant throughout
 Trials are made under same set of conditions
 BD – Probability density function
 Probability of getting ‘x’ Successes out of ‘n’
trials with ‘p’ probability of Success and ‘q’
probability of Failure is
P(X) = B(n,p,x) = ncxpxqn-x
where x = 0,1,2,..n
n
cx = n! / x!(n-x)!

 n and p are the two parameters of BD

 BD – Mean and Variance
 Mean of BD is the theoretical expected number of
successes in n trials
μ = E(X ) = np
 Variance of BD measures the variation of the
binomial
σ2 = npq
 Standard Deviation BD
σ = √npq
 BD: Problem 1
Output of a production process is known to be 30%
defective. What is the probability that a sample of 5
items would contain 0,1,2,3,4and 5 defectives?
 BD: Problem 1 - Solution
S = Defective item
‘p’ = 0.3 (30%);
‘q’ = 1 – p = 1 – 0.3 = 0.7
n=5 and P(X)= ncxpxqn-x
No of Successors(x) Probability P(x)
0 0.1681
1 0.3602
2 0.3087
3 0.1323
4 0.0283
5 0.0024
Total 1.0000
 BD : Problem 2
An incidence of occupational disease in an industry
is such that the workers have a 20% chance of
suffering from it. What is the probability that out of
six workers 4 or more will contract disease?
S = worker suffering from the disease
‘p’ = 0.2 (20%)
‘q’ = 1 – p = 1 – 0.2 = 0.8
n = 6 and P(x) = ncxpxqn-x
 BD : Problem 2 solution
The probability that 4 or more , i.e., 4,5,and 6 will
contract the disease is given by
P(X≥4) = P[4] + P[5] + P[6]
P[4] = 6 c 4 (1/5)4 (4/5)6-4
P[5] = 6 c 5 (1/5)5 (4/5)6-5
P[6] = 6 c 6 (1/5)6 (4/5)6-6
P(X≥4) = 0.01696
 Fitting Binomial Distribution
 If a random experiment consists of n trials, satisfying
the conditions of BD is repeated N times, then the
frequency of x successes is given by the formula
N * P (X=r) = N * (ncxpxqn-x ) where x = 0,1,2,..n
 The expected or theoretical frequencies of the BD are
obtained as given in the following table
Number of Successes(r) Expected or theoretical
frequencies N*P((X)
0 N*[qn]
1 N*[ nc1qn-1 p1]
2 N*[ nc2qn-2 p2]
. .
. .
n N*[ pn]
 BD: Problem 3
In an office there are 150 employees. The pattern of
their absence for duty in a particular month is
recorded in the following table. Fit a Binomial
Distribution to the data.

No.of days 0 1 2 3 4
absent:X
No.of 28 62 46 10 4
Absentees:
f(x)
 BD: Problem 3 solution
Steps :
(i) N = 4; N = ∑f = 150
(ii) Find Mean of the distribution ∑f(x) / ∑f
= (0+62+92+30+16)/150=200/150=4/3 = 1.33
(iii) μ = 4/3 or 1.33; np = 4/3 or 1.33;
hence, p = 1/3 or 0.33
and q = 1 – p = 2/3 or 0.67
(iv) Obtain P(x=0) to P(x = 4) using the formula
(ncxpxqn-x )
(v) Multiply these values with N = 150
 x P(x) Frequency
f(x)=N*P(x)=150*P(x)
0 4
c 0 (1/3)0(2/3)4=16/81= 0.1975 150*0.1975=29.63=30
1 4
c 1 (1/3)1(2/3)3=4(8/81)=0.3951 150*0.3951=59.26=59
2 4
c 2(1/3)2(2/3)2=43/21)(4/81)=0.2963 150*0.2963=44.44=44
3 4
c 3 (1/3)3(2/3)1 =4(2/81)=0.0988 150*0.0988=14.81=15
4 4
c 4 (1/3)4(2/3)0=1/81=0.123 150*0.0123=1.85=2

 Hence the fitted Binomial Distribution is

X 0 1 2 3 4 Tota
l
f(x) 30 59 44 15 2 150
POISSON DISTRIBUTION (PD)
 PD
 Also known as the “Law of Improbable
Events”, Poisson Distribution was derived in the
year 1837 by French Mathematician Simeon
Poisson
 It is used in cases where the chance of any
individual event being a success is small
 It is used to describe rare events like number of road
accidents, number of printing mistakes in a book,
etc.
 PD - conditions
Poisson Distribution may be obtained as a limiting
case of Binomial Distribution under the following
conditions:
i) n, number of trials is infinitely large, ie., n∞
ii) p, the constant probability of success for each
trial is infinitely small, i.e., p 0
iii) np=λ is finite
 PD - Examples
The following are some of the events which may be
analyzed using Poisson Distribution:
a) The demand for a product,
b) Typographical errors occurring on the pages of a book,
c) The occurrence of accident in a factory,
d) The arrival pattern of customers in a departmental store,
e) The occurrence of flaws in a bolt factory, and
f) The arrival of calls at a switch board.
PD – Probability density function
Under the above three conditions, the BD
Probability density function tends to PD
Probability density function as given below:
P(X) = e-λ λx/x! where x = 0,1,2……. ∞
(the number of successes)
λ = np and
e = 2.71828
 PD – Mean and Variance
 Mean of PD: µ = E(X) = ∑xf(x) = ∑x e-λ λx/x!
= 0+λ e-λ+(λ2/2!) e-λ+(λ3/3!) e-λ+..
= λ e-λ(1+ λ + λ2 /2!+ λ3 /3!...)
= λ e-λ eλ= λ
 Variance of PD: σ2 = E(X2) - [E(X)2 = E(X2)-λ2
= ∑x2 e-λ λx/x!- λ2
= λ2+ λ- λ2 = λ
Thus the variance of Poisson distribution is also λ.
 The Standard Deviation (SD) of PD = σ = √ λ
 PD: Problem 1
On an average, one in 400 items is defective. If the
items are packed in boxes of 100, what is the
probability that any given box of items will contain:
a) no defective
b)less than two defectives
c) one or more defectives
d) more than three defectives
 PD: Problem 1 solution
 Probability of a defective item p = 1/400;
number of items packed in the box n = 100;
λ = np = 100 x 1/400 = 0.25;

a) Probability of no defective
= P(X=0) = e-λ = e-0.25 = 0.7788

b) Probability of less than two defectives

= P(X<2) = P(X≤1)= P(X=0)+P(X=1)
= e-λ+ λ e-λ= e-λ(1+λ)
=0.7788(1+0.25)= 0.9735
 PD: Problem 1 solution
c) Probability of one or more defectives
P(X≥1)=1-P(X=0)= 1- e-λ= 1-0.7788 =0.2212

d) Probability of more than three defectives

P(X>3)=P(X ≥4) = 1-P(X ≤3)
= 1-[P(X=0)+P(X=1)+P(X=2)+P(X=3)]
=1-[e-λ +λ e-λ+(λ2/2) e-λ+(λ3/6) e-λ]
=1- e-λ[1+ λ +(λ2/2)+(λ3/6)]
=1-0.7788[1+0.25+(0.25)2 /2+(0.25)3 /3
=1-0.7788[1+0.25+0.03125+0.0026]
=1-0.7788[1.28385]
=1-0.99986=0.00014
 PD: Problem 2
A factory produces blades in packets of 10. the
probability of a blade to be defective is 0.2%. Find
the number of packets having two defective blades
in a consignment of 10,000 packets.
 PD: Problem 2 solution
 λ = np = 10 x 0.002 = 0.02
 P(X = 2 defective blades)
= e-0.02 (0.02)2/2
=(0.9802x0.0004)/2
= 0.4901 x 0.0004
= 0.000196
Therefore, the total number of packets having two defective
blades in a consignment of 10,000 packets is
10,000 x 0.000196 = 1.96 = 2
 Fitting PD
 Steps:
(i) Find mean of the given frequency
distribution ∑f(x) / ∑f .
(ii) Equate µ to λ and find Poisson
probabilities.
(iii) Multiply the probability values obtained
in the above step with N = ∑f to obtain Poisson
frequencies.
 PD: Problem 3
The following table gives the number of days in a
50-day period during which x automobile accidents
occurred in a city. Fit a Poisson distribution to the
data:

No. of accidents: 0 1 2 3 4
No. of days: 21 18 7 3 1
 PD: Problem 3 solution
 Steps:
(i) Mean of the given frequency distribution is
0 + 18 + 14 + 9 + 4 / 50 = 0.9
(ii) Hence, λ = 0.9. Poisson probabilities are
f(0)= e-λ= e-0.9 =0.4066,
f(1)= λf(0)=0.9 x 0.4066=0.3659
f(2)= (λ/2)f(1)=(0.9/2) x 0.3659=0.1647
f(3)= (λ/3)f(2)=(0.9/3) x 0.1647 = 0.0494
f(4)= (λ/4) f(3)= (0.9/4) x 0.0494 = 0.0111
 PD: Problem 3 solution

(iii) Multiply each probability by N to fit PD

X Poisson N*P(x)
valu Probability
e P(x)
0 0.4066 0.4066 * 50 = 20.33 ~ 20
1 0.3659 0.3659 * 50 = 18.33 ~ 18
2 0.1647 0.1647 * 50 = 8.24 ~ 8
3 0.0494 0.0494 * 50 = 2.47 ~ 2
4 0.0111 0.0111 * 50 = 0.56 ~ 1
NORMAL DISTRIBUTION (ND)
 ND
 Also known as Laplace , Gauss and famous as
Normal distribution, it was discovered by De
Moivre as the limiting case of Binomial model in
1733.
 It is approximation of BD as it tends to the form of
the continuous curve when n becomes large
 The limiting frequency curve, obtained as n,
becomes large and is called the Normal frequency
curve or simply the Normal curve
 ND – Probability density function
 A random variable X is said to have a Normal Distribution with
parameters μ (mean) and σ2 (Variance) if the density function is
given by

1
y= f(x) = ---- e -1/2{x-μ/ σ}2 ; -∞< x <+∞
σ√2∏

Where y is the computed height of an ordinate at a distance of x units from

he mean μ
∏= 22/7= 3.1416 (is a constant)
e= 2.7183 (the base of the system of natural logarithms and is constant
μ =Mean and
σ2 =Variance of the given random variable X
 ND – Probability density function
Expressedsymbolically as X~N(μ ,σ2 ) and read as :
Random Variable ‘X’ follows
Normal Distribution with mean ‘μ’ and Variance
‘σ2’
 ND - Graph
 Graphical representation of ND is known as Normal
curve and is given below:

 The graph of y = f(x) is a famous ‘bell shaped’

curve
 The top of the bell is directly above the mean μ
 For large values of μ, the curve tends to flatten out
and for small values of σ2, it has a sharp peak
 Standard Normal Distribution
 A Random Variable with any mean and standard
deviation can be transformed to a ‘Standard Normal
Variate’ (SNV) by subtracting mean and dividing by
standard deviation.
 For a Normal Distribution with mean μ and standard
deviation σ the SNV’Z’ is obtained as Z = x - μ / σ
 The value of Z represents the distance expressed as
a multiple of the standard deviation that the value X
lies away from mean.
 Standard Normal Distribution
 SNV Z has mean zero and variance one.
 Symbolically, if X~ N ( μ,σ2 ), then Z ~ N(0,1)
[If X follows ND with mean μ and variance σ2 then Z follows ND
with mean o and variance 1]
 The probability density function of SNV z is given
by y = f(z) =1/√2∏*e1/2 (z)2; -∞ < z < ∞
 Properties of ND
 As distinguished from Bionomial and Poision
Distributions where the RV is discrete, the RV
associated with the Normal curve is a continuous
one
 The normal curve is symmetrical about the mean
(skewness=0) . If the curve is folded along its
vertical axis, the two halves will coincide
 Properties of ND
 The number of observations below the mean are
equal to the number of observations above the
mean, which makes the mean and median coincide
 First and third qualities are equidistant from the
median
 The maximum height (Mode) of the normal curve is
at its mean. Thus for a ND mean, median and mode
are equal.
 Properties of ND
 There is only one maximum point for the Normal
curve which occurs at the mean. ND is a uni-modal
distribution
 The curve extends on either sides of mean
approaching X-axis but it never touches X axis. This
is known as Asymptotic property of the Normal
curve. Its range is unlimited or infinite on both
directions
 Mean deviation about mean is (4/5) s or more
precisely 0.7979 times of standard deviation
 Area property of Normal Curve
The areas covered under normal curve on either side
of the mean are tabulated below.
Limits Area in %
μ ± 0.5σ 38.30
μ ± 1.0σ 68.26
μ ± 1.5σ 86.64
μ ± 2.0σ 95.44
μ ± 2.5σ 98.76
μ ± 3.0σ 99.74
μ ± 3.5 σ 99.96
Area property of Normal Curve
Calculating areas under the ND curve
 Areas under a Normal curve are calculated by using
SNV tables where Z~N(0,1)
 Since the Normal curve is symmetrical, the area
under the Normal curve for a negative value is same
as that of the area under the normal curve for a
positive value
 Calculating areas under the ND curve
AREA CORRESPONDING TO Z = 1 AREA CORRESPONDING TO Z = - 1
UNDER THE NORMAL CURVE IS UNDER THE NORMAL CURVE IS
GIVEN AS 0.3413 (FROM THE ALSO 0.3413 BECAUSE OF
TABLE) SYMMETRIC PROPERTY OF ND
 ND: Problem 1
Since this area cannot be
read directly from the
Find the area under normal tables, first read the area
curve above Z = 0.25. Between Z=0 and Z=
0.25from the table as
0.0987
 ND: Problem 1 solution
 The required area is
obtained as
0.5-0.0987=0.4013

 Hence the area under the

Normal curve above
Z=0.25 is 0.4013
 ND: Problem 2
If the salary of workers in a factory is assumed to
follow a Normal Distribution with a mean of Rs.500
and a S.D.of Rs.100 find number of workers whose
salary is between Rs.400 and Rs.650, given the
number of workers in the factory as 15000.
 ND: Problem 2 solution
 Calculate Z values corresponding to the given X
values
(i) z1= (x1-µ)/σ = (400-500)/100 = -1
(left of mean)
(ii) z2= (x2-µ)/σ =(650-500)/100= +1.5
(right of the mean)
 ND: Problem 2 solution
 Find the corresponding areas from the table
values
(i) Area between Z = 0 and Z = -1 is 0.3413
(ii) Area between Z = 0 and Z = 1.5 is
0.4332
 P(400<X<650) is 0.3413+0.4332 = 0.7745
 No. Of employees drawing salary between
Rs.400 and Rs.650 is 15000 * 0.7745 = 11618