Random Variable

Basic Preliminaries: -
 Random Experiment::Experiment whose outcomes are not predictable.
 Sample Point:: Each and every outcome of random experiment.
 Sample Space:: Totality or aggregate of sample points. It is denoted by symbol S or 
 Event:: A subset of sample space.
 Impossible Event:: Event which does not contain any sample point.
 Certain Event::Event which contains all sample points.
 Mutually Exclusive events:: The happening of any one event excludes the happening of
the other event.
n
 Exhaustive Event::The events  A1 , A2 , , An  are said to be exhaustive if Ai  S .
i 1

 Independent Events:: The occurrence or non occurrence of one does not affect the
occurrence or non occurrence of the other.
# Probability
If a random experiment results in ‘n’ mutually exclusive and equally likely outcomes of
which ‘m’ are favorable to event A then probability of event A, denoted as
m
p ( A) and is defined as p ( A)  .
n
Standard Results ::
1. 0  p( A)  1
2.  
p  AC   p A 1  p  A
3. p    0, p  S  1
4. If A  B then p ( A)  p ( B )
5. p  A  B   p  A  p  B   p  A  B 
6. If A and B are mutually exclusive events then p  A  B   0
7. In particular, A and B are mutually exclusive p  A  B   p  A  p  B 
p  A  B p  A  B
8. Conditional Probability: p  A / B   , p  B / A  .
p  B p  A
Thus p  A  B   p  A p  B / A , p  A  B   p  B  p  A / B 
For n events A1 , A 2 , , A n the probability of intersection event is
p  A1  A2  A3   An   p  A1  p  A2 / A1  p  A3 / A1  A2  p  An / A1  A2  An1 
Sufficient condition for above result to hold is p  A1  A2  An1   0 .
9. If A and B are independent events then p  A  B   p  A p  B  .
n
10.  A1 , A2 , , An  is a set of mutually exclusive events then  p  Ai   p  S   1 .
i 1
11. Binomial Probability :: An experiment is performed n number of times repeatedly. A is
an event known as success with probability p. If event A occurs r times among n trials
then P(r successes )  nCr p r q n  r , where p is the probability of success and q  1  p the
probability of failure.
 Random Variable
Random Variables
Let  be a sample space associated with given random experiment. A function defined
from  to a number set is known as Random Variable. Thus a random variable is a function
whose domain set is a sample space of some random experiment.
The collection of values actually taken by a random variable is known as Range of the random
variable. Depend on the range; random variables get classified into the following two types.
 Discrete Random Variables:: Random variable whose range set is a discrete set or the
set of natural numbers is known as discrete random variable.
e.g. Number of accidents on Mumbai-Pune expressway, Number of Neutrons emitted by a
radioactive isotope, Number of students in a class.
 Continuous Random Variables:: Random variable whose range set is an interval is
known as continuous random variable.
e.g. Lifetime of an electrical or electronic components, Height of students, Amount of cosmic
radiation, Time between arrivals of two flights at certain airport, Time taken for surgery at a
certain hospital.
Note : Random variables arising through the counting processes are usually discrete in nature
while those which arises through measuring processes are continuous in nature.
Examples of Random Variables
1. Consider the experiment of tossing of two fair coins simultaneously. The sample space is
  HH , HT , TH , TT  . Let X 1  count of heads in a single toss. By this every sample point get
assign by a real number viz X1 ( HH )  2, X1 ( HT )  X1 (TH )  1, X1 ( HH )  0 .
Thus X1 is a random variable defined on  . Here Range of X1  0,1, 2 . Since the range of
X1 is a subset of discrete set, X1 is a discrete random variable.
Similarly, X 2  count of tails in a single toss is also a discrete random variable defined on  .
2. Consider the experiment of tossing of two fair dice simultaneously.
Sample space   (i, j ) /1  i, j  6 .
i) X 1  Sum of the numbers appear on the two faces, i.e., X1 ((i, j ))  i  j .
Range of X1  2,3, 4,5,6,7,8,9,10,11,12 .
ii) X 2  Product of the numbers appear on the two faces, i.e., X 2 ((i, j ))  ij
Range of X 2  1, 2,3, ,36 .
iii) X 3  Minimum or Maximum of the two numbers appear on the two faces, i.e.,
X 3 ((i, j ))  min(i, j ) or max(i, j ) . Range of X 3  1, 2,3, 4,5,6 .
Note : Range of all X1 , X 2 and X 3 is a subset of set of Natural number set(Discrete set)
Hence X1 , X 2 and X 3 are discrete random variables (drv).
iv) X 4  Quotient of the two numbers appear on the two faces, i.e., X 4 ((i, j ))  i / j . Range
of . Here X 4 is taking integer as well as rational values.

Therefore range of X 4 is an interval . Hence X 4 is a continuous random variable

(crv).
v) Reciprocal of the sum of two numbers appear on the two faces, i.e., .
 Random Variable
Range of . Here is taking integer as well as rational values.

Therefore range of is an interval Hence is a continuous random variable (crv).

Probability Mass Function (p. m. f.)
The probability mass function of discrete random variable X with range set
x1, x2 , , xn  defined on a sample space  is the assignment pi  p( xi )  p( X  xi ) such that
n
i) p( xi )  0  i  1, 2,3, , n. ii)  p( x )  1 .
i 1
i

The table containing the value of X along with the probabilities given by probability
mass function is called probability distribution of the random variable X .
Examples
1. Consider the experiment of tossing of 3 coins simultaneously.
  HHH , HHT , HTH , THH , HTT , TTH , THT , TTT  .
Let X be count of heads. Then c. d. f of X can be tabulated as follows.
X  xi 0 1 2 3
p( X  xi ) 1/ 8 3/ 8 3/ 8 1/ 8
2. Consider the experiment of tossing of 2 fair dice simultaneously.
  (i, j ) /1  i, j  6 . Let X be count of sum of the numbers appear on the
faces.
X  xi 2 3 4 5 6 7 8 9 10 11 12
p( X  xi ) 1 2 3 4 5 6 5 4 3 2 1
36 36 36 36 36 36 36 36 36 36 36
3. A point is chosen at random in a circle of radius . Let X be the distance of the point
from the centre of the circle. Then the range of X is the closed interval with end points
and , i.e., range is Here X is a continuous random variable.
Distribution Function or Cumulative Distribution Function (c. d. f )
Let be a discrete random variable with rangeset .The distribution function
of X denoted as , is the probability of the event , i. e.,

Let X be a discrete random variable taking values  x1 , x2 , , xn  with the pmf

X  xi x1 x2 x3 xn
p( X  xi ) p1 p2 p3 pn

Then the cumulative distribution function (cdf) of X is

X  xi x1 x2 x3 xn
p( X  xi ) p1 p2 p3 pn

p1
 Random Variable

Note that: i) Probability Mass Function (pmf) is a function of discrete variable. There are two
ways for graphical representation of pmf. The bar chart and the histogram. The sum of the
lengths of the bars in the bar chart is 1 whereas the sum of the areas of the rectangles in the
histogram is 1.
ii) is a function of real continuous variable . Graph of this function is step or staircase.
Graphs of pmf and cdf :

Illustrative Examples
Q 1) If X is a random variable the difference between heads and tails obtained when a fair
coin is tossed 3 times. What are the possible values of X and its probability mass
function? Also write the distribution function of X .
Sol .   HHH , HHT , HTH , THH , HTT , TTH , THT , TTT  . X can take values from
n

-3 to 3, i.e., Range of X  3,  1,1,3 . The pmf is

X  xi 3 1 1 3
1 3 3 1
p( X  xi )
8 8 8 8
1
8

Q 2) A fair dice is rolled twice. Find the possible values of random variable X and its
associated probability mass function, where X is the maximum of the values appearing
in 2 rolls.
n.   (i, j ) /1  i, j  6 . X can take values from 1 to 6, i.e.,
Sol
Range of X  1, 2,3, 4,5,6 . The pmf is
X  xi 1 2 3 4 5 6
1 3 5 7 9 11
p( X  xi )
36 36 36 36 36 36
Q 3) A random variable X takes values 3,  1, 2, 5 with respective probabilities
2k  3 k  1 k  1 k 2
, , and . Determine the distribution of X .
10 10 10 10
Soln.Theassignments are probabilities, equating the to one
2k  3 k  1 k  1 k  2
    1  k  3 . Hence the distribution of X is
10 10 10 10
X 3 1 2 5
p( X  x) 3/10 4 /10 2 /10 1/10
 Random Variable
Q 4) A random variable X has probability mass function (pmf) shown in the following
tabular form. Find the value of unknown k . Hence write pmf and cdfof X . Draw
graphs of pmf and cdf . Also find i) p (1  X  3) ii) p (1  X  3)
iii) p( X  1) iv) p ( X  5)

X 1 2 3 4 5
p( X  x) k / 36 3k / 36 4k / 36 10k / 36 18k / 36

Soln.Since the above assignment is probability distribution, sum of the probabilities

is one implies k  1. Thus the pmf and cdf of X are
X 1 2 3 4 5
p( X  x) 1/ 36 3/ 36 4 / 36 10 / 36 18 / 36
FX ( x) 1/ 36 4 / 36 8 / 36 18 / 36 36 / 36 1

1 3 4 1
i) p 1  X  3  p  X  1  p  X  2     .
36 36 36 9
3 4 7
ii) p 1  X  3  p  X  2   p  X  3    .
36 36 36
1. Mathematical Expectation / Theoretical Mean (analogous to Centre of Gravity)
Theoretical mean or expectation of X denoted as E ( X ) or  is defined as
n
E ( X )   xi pi .
i 1
Expectation value of X provides a central point of the distribution.
Note: Expected value of a random variable may not be actually taken by the variable.
2. Variance
Variance of X denoted as Var ( X ) .
n
Var ( X )    xi  E ( X )  pi . This can be simplified as
2

i 1
n
Var ( X )  E ( X 2 )   E ( X )  , where E ( X 2 )   xi 2 pi
2

i 1

3. Standard Deviation sd   Var ( X )

Results : Let X and Y be two random variables. Let a and b be any non zero constants.
i) E (aX  b)  aE ( X )  b
ii) E ( X  Y )  E ( X )  E (Y )
iii) Var (aX  b)  a 2 Var ( X )
iv) sd (aX  b)  a sd ( X )
Illustrative Examples
Q 1) A random variable X has the following p.m.f.
X 0 1 2 3 4 5 6
p( X  xi ) k 3k 5k 7k 9k 11k 13k

Find (i) k (ii) p  X  2 (iii) p  0  X  5

 Random Variable
(iv) What is the minimum value of C for which p  X  c   0.5
(v) What is distribution function of X ?
1
Soln. (i) i p  X  xi  1 k  49 .
45
(ii) p  X  2   1  p  X  2   1  p  X  0   p  X  1 
49
4
15
(iii) p  0  X  5    p  X  xi   .
i 1 49
16 25
(iv) p  X  3   0.5 and p  X  4    0.51
49 49
Minimum C is 4.
(v) p.m.f.
X  xi 0 1 2 3 4 5 6

p  X  xi 
1 3 5 7 9 11 13
49 49 49 49 49 49 49
c.d.f. F  a   p  X  a 
X  xi 0 1 2 3 4 5 6
1 4 9 16 25 36
F (a) 1
49 49 49 49 49 49
Q 2) Determine k such that the following functions are p.m.f.s
Soln. 1. P ( x )  k x, x  1, 2, 3, , 10
k (1  2  3  4  5  6  7  8  9  10)  1
1
k 
55
2x
2. P( x)  k , x  0, 1, 2, 3
x!
 4 3
k 1  2  2    1 k 
 3 19
3. P( x)  k (2 x 2  3x  1), x  0, 1, 2, 3
1
k (1  6  15  28)  1 k 
50
n
Q 3) Verify whether the assignment p( X  n)  2 , n  1, 2,3, is a probability mass
function for random variable X .
Soln. To be a probability mass function
i) p ( X  n)  0  n

ii)  p ( X  n)  1
n 1

1
p ( X  n)  is in exponential function, condition i) holds.
2n
   n
1 1 1
Further  p( X  n)   n      1  1  1
n 1 n 1 2 n 0  2  1  (1/ 2)
 Random Variable
Hence condition ii) also holds.Therefore the assignment is a probability mass function.
Q 4) A box contains 8 items of which 2 are defective. A person draws 3 items from
the box. Determine the expected number of defective items he has drawn.
n
Sol . Let X be the number of defective items drawn by a person. The pmf of X is
0 1 2

The expected number of defects .

1  3x 1  x 1  2 x
Q 5) A random variable X take four values with probabilities viz , , and
4 4 4
1 4x
. Find the condition on X that these values represent the probability mass function.
4
1  3x 1 x 1 2x 1 4x
Soln. Let p1  , p2  , p3  and p4  .
4 4 4 4
As this is pmf , 0  pi  1 for i  1, 2,3, 4 and p1  p2  p3  p4  1 .
All these probabilities sum up to 1. Hence consider first condition.
1  3x 1 1 x
0 1   x 1 , 0   1  3  x  1
4 3 4
1 2x 1 3 1 4x 3 1
0 1   x  , 0  1   x  .
4 2 2 4 4 4
1 1
Therefore   X  .
3 4
1
Q 6) A random variable has mean 2 and standard deviation . Find
2
i) E (2 X  1)
ii) Var ( X  2)
 3X 1 
iii) sd  
 4 
1 1
Soln. Given E ( X )  2 and sd ( X )   Va r( X ) 
2 4
i) E (2 X  1)  2 E ( X )  1  2* 2  1  3 .
1
ii) Var ( X  2)  Var ( X )  .
4
 3X 1  3 3 1 3
iii) sd   sd ( X )  *  .
 4  4 4 2 8
Q 7) A sample space of size 3 is selected at random from a box containing 12 items of which 3
are defective. Let X denote the number of defective items in the sample. Write the
probability mass function and distribution function of X. Find the expected number of
defective items.
Soln. X be the number of defective items in the sample.
 Random Variable
0 1 2 3
84/220 108/220 27/220 1/220
84/220 192/220 219/220 220/220=1

The expected number of defective items in a sample is

Q 8) A player tosses two fair coins. The player wins $2 if two heads occur, and $1 if one head
occur. On the other hand, the player losses $3 if no heads occur. Find the expected gain of
the player. Is the game fair?
Soln.The sample space is HH , HT , TH , TT  . Since coins are fair,
1
p  HH   p  HT   p TH   p TT   .
4
Let X be the player’s gain. Then X takes values 3, 1 and 2 with
1 2 1
p  3  , p 1  and p  2   .
4 4 4
1 2 1 1
Expected gain E  X   3   1  2    0.25 .
4 4 4 4
Thus the expected gain of the player is $ 0.25 . Further E  X   0 , the game is
Favorable to the player.
 Continuous Random Variable
Probability density function (p.d.f.)
X a continuous random variable. Function f ( x ) defined for all real x  (, )
is called probability density function (p.d.f.) if for any set B of real numbers, we get
probability p( X  B)   f ( x)dx .
B
b
Thus, p  a  X  b    f ( x)dx .Also,the point probability is zero, i.e., p  X  a   0 .
a

Axioms
b
i) 0  p  a  X  b   1, or , 0   f ( x)dx 1
a

Note that p( X  B)   f ( x)dx .

B

ii) 

f ( x)dx  1 .

Let f ( x ) be the probability density function (pdf) of a continuous random variable X .

The curve represented by the equation y  f ( x) is called the probability curve of X .
Note That :i) The probability curve lies completely above X - axis.
ii) The total area between the X - axis and the probability curve is unity.
iii) p  a  X  b  represents the area under the probability curve bounded by the X -
axis and the co-ordinates X  a and X  b .
 Random Variable

p  a  X  b

 Cumulative Distribution Function (c.d.f.)

X be a continuous random variable.
The cumulative distribution function denoted as F (a) and is expressed as
a
d
F (a)  

f ( x)dx or
da
f (a )  f (a ) .

Properties of c.d.f.
1) F (a) is monotonically increasing. i.e. a  b  F (a)  F (b)
2) lim F (a)  1
a 

3) lim F (a)  0
a 

4) p  a  X  b   F (b)  F (a)
Note that : For a continuous random variable X ,
p  a  X  b  p  a  X  b  p  a  X  b  p a  X  b 
5) p  X  a  1 p  X  a   1 F  a 

 Expected value, Variance and Standard Deviation

 
  E( X )   x f ( x)dx , var( X )   ( x  E ( X )) f ( x)dx, sd ( X )   Var ( X )
2

 
Illustrative Examples

Q 1) X cont. random variable with p.d.f

0 if x0
x

F ( x)   if 0 x2
 2
1 if 2 x
1 3
Find: i) p   X   (ii) p 1  X  2
2 2
1 3 3 1 3 1 1
Soln. i) p  a  X  b   F b   F  a  p  X   F      .
2 2 2 2 4 4 2

1 1
ii) p 1  X  2   F  2   F 1 1   .
2 2
 Random Variable
Q 2) The time in years X required to complete a software project has a p.d.f.
 kx(1  x), 0  x 1
f ( x)  
0 otherwise
Compute the probability that the project will be completed in less than four months.

Soln. 

f ( x)dx  1  k  6

1
The probability that the project will be completed in  4 months, i.e. less than
3
years.
1/3
p  X  4    f ( x) dx  0.259 .
0

Q 3) X continuousrandom variable with probability density function.

x
 , 0 x2
f ( x)   2
0, otherwise
Find (i) p 1  X 1.5 (ii) E ( X ) (iii) Var( X ) (iv) c.d.f.
Soln. This is a ramp distribution.
1.5
5
(i) p 1  X 1.5   f ( x) dx  16
1
2
x
2
4
(ii) E ( X )   dx 
 2 3
0
2
(iii) E ( X 2 )  2  Var( X ) 
9
a
(iv) F (a)  

f ( x)dx
0
  x  0  0dx  0

a
2
0 x2  x dx  a

2 4
0

2 x F (a)  1
Q 4) X continuous random variable with probability density function
kx, 0 x2

f ( x )   2k , 2 x4
kx  6k , 4 x6

Find k and mean.
6 2 4 6
Soln. 
0
f ( x) dx  1  kx dx   2k dx   (kx  6k ) dx  1
0 2 4
 Random Variable
1
Evaluating the integral and simplifying, we have 8k  1  k  .
8
6 2 4 6
E ( X )   xf ( x) dx   kx 2 dx   2kx dx   (kx 2  6kx) dx
0 0 2 4
2 4 6 6
 x3   x2   x3   x2 
 k    2 k    k    6k  
 3 0  2 2  3 4  2 4
Simplifying E  X   3 .
Q 5) i) Verify f ( x)  6 x 1  x  ,0  x  1 is a probability density function for the
diameter of a cable.
ii) Find the condition on b such that p  X  b   p  X  b  .
Soln.i) At x  0, f ( x)  0. For 0  x  1 , 0  1  x  1and 0  6x  6 . Therefore
0  6 x(1  x)  6 . Hence f ( x ) is positive for 0  x  1 .
1 1
1
Further  f ( x)dx  0 6 x(1  x)dx  6  2  3   1 .
This proves f ( x ) is pdf for 0  x  1 .
ii) p  X  b   p  X  b   p  X  b   1  p  X  b  .
Random variable is continuous, p  X  b   p  X  b  . Therefore
b
1 1
2 p  X  b   1  p  X  b     f ( x)dx  .
2 0 2
Solving above equation we have 4b3  6b 2  1  0.
Q 6) A continuous random variable X has pdf f ( x)  3x 2 , 0  x  1 . Find a such that
p  X  a   0.05 .
1
Soln. p  X  a   0.05   f ( x)dx  0.05 , i.e.,
a
1

 3x dx  0.05  1  a 3  0.05  a   0.95  .

2 1/3

 Probability Models or Statistical Distributions

In many practical situations, the random variable follows a specific pattern which can be
described by a standard probability distribution. In these cases, probability mass function can be
expressed in algebraic form and various characteristics of the distribution can be calculated using
known closed formulae.
Bernoulli Distribution
Bernoulli random variableis an outcome of Bernoulli trial. Suppose, an experiment whose
outcome can be classified as either a success or a failure. The probability of success is p and
failure is 1  p . Then the Bernoulli random variable X is defined as X  1 if trial is a success
and X  0 if trial is a failure.
The distribution is discovered by Swiss Mathematician James Bernoulli, in 1713.
It is applied wherever the experiment results in only two outcomes.
 Random Variable
One outcome is termed as a success 1 and the other as failure 0. Such an experiment is called
Bernoulli Trial.
 Probability Mass Function
 pi (1  p)1i i  0, 1
P( X  i )  
0 otherwise
random variable X  xi 0 1
probability mass function 1 p p
 Cumulative Distribution Function
0 a0

F (a)  1  p 0  a  1
1 1 a

 Expectation / Mean
E ( X )   i p(i)  0(1  p)  1 p  p
i

 Variance
Var ( X )  E ( X 2 )  ( E ( X )) 2   i 2 p(i)  p 2  p  p 2  p(1  p)
i

 Applications
Situations where Bernoulli distribution is commonly used are
1) Items produced by a machine defective / non-defective.
2) Students appearing for examination pass / fail.

Binomial Distribution
Binomial random variable counts the number of successes when n Bernoulli trials are
performed.Suppose n independent Bernoulli trials are performed. Each trial results in a success
with probability p . If X represents the number of successes that occur in n trials, then X is
said to be a Binomial random variable and the probability distribution is known as Binomial
distribution with parameters (n, p ) .
 Probability Mass Function (p.m.f.)
probability mass function of a binomial random variable having parameters (n, p ) is given
by P( X  i )  nCi p i (1  p ) n i , i  0, 1, , n .
n
Note that  nCi p i (1  p ) n i  ( p  1  p ) n  1 .
i 0

 Cumulative Distribution Function (c.d.f.)

a
F (a )  P ( X  a )   nCi p i (1  p ) n i
i 1

 Expectation
n n
E ( X )   i p(i )   i nCi p i (1  p ) n i
i 0 i 1

n! n! n(n  1)! n(n  1)!

i nCi  i      n n 1Ci 1
i !(n  i )! (i  1)!(n  i )! (i  1)!(n  i)! (i  1)!(( n  1)  (i  1))!
 Random Variable
n n
 E ( X )   n n 1Ci 1 p i (1  p ) n i  n  n 1Ci 1 p i (1  p ) n i , Put i  1  r
i 1 i 1
n 1 n 1
 E ( X )  n n 1Cr p r 1 (1  p) n  r 1  np  n 1Cr p r (1  p) n 1r  np( p  1  p ) n 1  np
r 0 r 0

 Variance
n n n
E ( X 2 )   i 2 p(i )   i 2 nCi p i (1  p) n i   i 2 nCi p i (1  p) n i
i 0 i 0 i 1
n 1 n 1

r
r 0
n 1
Cr p r (1  p ) n 1 r  n n  2Cr 1 p r (1  p ) n 1r
r 0
n n n
E ( X )   (i  i  i ) Ci p (1  p )
2 2 n i n i
  i (i  1) Ci p (1  p )
n i n i
  i nCi p i (1  p ) n i
i 1 i 1 i 1

i (i  1) n ! n! (n  2)!
Now, i (i  1) nCi    n(n  1)  n(n  1) n  2Ci  2
i !(n  i )! (i  2)!(n  i)! (i  2)!(( n  2)  (i  2))!
n n
  n(n  1) n2
Ci  2 p (1  p)
i n i
 n(n  1) n  2Ci  2 p i (1  p) n i
i 1 i 2
n2 n2
i  2  r  n(n  1) n  2Cr p r  2 (1  p) n r  2  p 2 n (n  1) n  2Cr p r (1  p )( n  2)  r
i 0 i 0
n2
 p n (n  1)( p  1  p)
2
 p n (n  1)
2

 E ( X )  p n (n  1)  np  p 2 n2  p n n  np  p 2n2  np(1  p)
2 2

 Var ( X )  E ( X 2 )  ( E( X ))2  { p 2 n2  np(1  p)}  n2 p 2  np(1  p)  npq

Therefore sd ( X )  npq .
Mode of the distribution  (n  1) p   integral part of ( n  1) p .
 Applications
Binomial distribution can be used when
1. Trials are finite, performed repeatedly n times.
2. Each trial should result as success or failure.
3. Probability of success is constant for each trial as p .
4. All trials are independent.

Illustrative Examples

Q 1) The probability that a bomb dropped from a plane will strike the target is the probability
that (i) exactly two will strike the target (ii) at least two will strike the target.
Soln. Random Variable X : Count of bombs dropped from a plane striking the target
1  1
p  & n  6  X B  6,  .
5  5
(i) Exactly two will strike the target.
4444
2 4
1  4
P( X  2)  C2      15  (0.2)6  153 
6
 0.24576
5  5 555555
(ii) At least two will strike the target.
P( X  2)  1  P( X  2)
 Random Variable
 1  P( X  0)  P( X  1)
0 6 5
1 4  1  4 
 1       6    
5  5  5  5 
 4  4  4  4  4  4 6  45 
 1   6 
 5 5 5 5 5 5 5 
45 45
 1  6 [4  6]  1  102   1  0.65536  0.34464
5 5  55
Q 2) The probability that an entering student will graduate is 0.4. Determine the probability
that out of 5 students (a) none (b) one (c) at least one will graduate.
Soln. Random Variable X : Count of graduate students p  0.4, n  5
 X B (5, 0.4) .
(a) P( X  0)  (0.4)0 (0.6)5  0.07776
(b) P( X  1)  0.2592
(c) P( X  1)  1  P( X  0)  0.92224

Q 3) Point out the fallacy of the statement : The mean of binomial distribution is 3 and
variance is 5.
5
Soln. np  3 & npq  5 q   1
3
Q 4) The mean and variance are 6 and 2 respectively. Find P ( X  3) .
1
Soln. np  6 & npq  2  q 
3
1 2
p  1  nq
3 3
P( X  3)  1  P( X  3)
 1  P( X  0)  P( X  1)  P( X  2)
 2  1  9  8  2   1 
0 9 8 2 7
 2 1
 1       9         
 3  3  3  3  2  3  3
9
1
P( X  1)  1  P( X  1)  1  P( X  0)  1     0.999949
3
Q 5) If the probability that a new born child is a male is 0.6, find the probability that in a
family of 5 children there are exactly 3 boys.
Sol . X : count of a new born male child. Given p  0.6 & n  5  X B(5, 0.6) .
n

5!
P( X  3)  5C3 (0.6)3 (1  0.6) 2  (0.6)3 (0.4) 2  10  (0.6)3  (0.4) 2  0.3456 .
3! 2!
Q 6) Out of 800 families with 5 children each, how many would you expect to have
i) 3 boys?
ii) either 2 or 3 boys?
Assume equal probabilities for boys and girls.
1  1
Soln. Let X = Number of boys. Given p  , n  5 . X B  5,  .
2  2
 Random Variable
3 2
1 1
i) p ( X  3)  5C3      0.3125
2 2
 Total number of families with 3 boys = 0.3125*800  250
3 2 2 3
1 1 1 1
ii) p( X  2 or X  3)  5C3      5C2      0.625
2 2 2 2
 Total number of families with either 2 or 3 boys = 0.625*800  500 .

Poisson Distribution
A random variable X said to follow Poisson distribution if its probability mass function
e   r
is given p( X  r )  , r  0,1, 2,3 .  is known as parameter of the distribution.
r!
 Expectation and Variance
If X P ( ) then E ( X )   , Var ( X )   sd ( X )   .

Poisson distribution as a limiting form of Binomial distribution

Poisson distribution can be derived as a limiting case of the Binomial distribution.
X B(n, p ) if
i) The number of trials n is large, i.e., n  
ii) Probability of success is very small , i. e., p  0
iii) With this quantity   np is finite.
 Applications
Poisson distribution can be used when
i) The rate of occurrence is constant over the time or space
ii) Past occurrences do not influence the likelihood of future occurrences.
iii) Simultaneous occurrences nearly impossible.
Following are few situations where Poisson distribution is applicable.
i) Arrival pattern of ‘defective’ items in a factory
ii) Arrival pattern of ‘patients’ in a hospital.
iii) Demand pattern of certain items
iv) Spatial distribution of bomb hits.

Illustrative Examples

Q 1) For a Poisson random variable X , p( X  1)  2 P( X  2) . Find the parameter 

of the distribution. Also find p ( X  2) .
e   e   2
Soln. p( X  1)  2 P( X  2)    2
1! 2!
p( X  2)  1  p( X  2)  1   p( X  0)  p( X  1) 
 1   e2  2e2   1  3e2
Q 2) In a certain factory turning out razor blades, there is a small chance of 0.002 for
any blade to be defective. The blades are supplied in packets of 10, use Poisson
distribution to calculate the approximate number of packets containing no
defective. One defective, two defective blades respectively in a consignment of
 Random Variable
10000 packets.
Soln. X : the blade to be defective.
p  0.002, n  10,   np  10  0.002  0.02
Taking X as a Poisson r.v.
e0.02  (0.02)0
i) P( X  0)   0.980198673
0!
 No. of packets containing no defective blades
 0.98019867310000  9801.98673
0.02
e  (0.02)
ii) P( X  1)   0.019603973
1
e0.02  (0.02) 2
iii) P( X  2)   1.960397346 104
2
 No. of blades  1.960397346
Q 3) A manufacturer of cottor pins knows that 3% of his product is defective. If he sells
cottor pins in boxes of 100 pins and guarantees that more than 5 pins will be
defective in a box, find the approximate probability that a box will fail to meet the
guaranteed quality.
Soln. X : product is defective. p  0.03, n  100,   3
P( X  5)  1  P( X  5)
5
e3  3
 1 
r 0 r!
 1  0.10081813  0.049787068  0.149361204
0.02240418  0.224041806  0.340263493
 0.113323436
 Normal Distribution
Any quantity whose variation depends on random causes is distributed according to the
normal law. A large number of distributions approximate to the normal distribution.
Discovered by De Moivre in 1733.
Let X be a binomial variable with mean np and standard deviation npq .
X  np
Define Z ---------- (i)
npq
Then Z is a binomial r.v. with mean 0 and standard deviation 1.
As n  , Z can take values between  to  .
In limiting form as n   when neither p or q is very small, Z follows a normal distribution.
A normal r.v. Z is a continuous r.v.
The probability density function of a given random variable is given by
( x  )2
1 
f ( x)  e 2 .
2

 2
where,   mean and   standard deviation.
x 
Note that Z  then E ( Z )  0 & Var( Z )  1 .

In this case, Z (0, 1) is known as standard normal variate whose pdf is
 Random Variable
1  z 2 /2
f ( z)  e
2
 Properties of normal distribution
1. The normal curve is bell-shaped and symmetric about its mean.

2. The probability of x lying between x1 & x2 is given by the area under the normal curve
from x1 to x2 .
z
1
In terms of standard normal variate, p(0  z )   e  z /2 dz .
2

2 0
This integral is called the probability integral or the error function.

Note that:
i) The area under the normal curve between the ordinates x     &
2
x     is 0.6826, i.e. 68.5%.  Approximately of the values lie within
3
these limits.
ii) The area under the normal curve between x    2 & x    2 is 0.9544 nearly
1
95.5% which implies that 4 % of values lie inside these limits.
2
iii) 99.73% values lie inside x    3 to x    3 .
iv) 95% values lie inside x    1.96 to x    1.96 , i.e. only 5% lie outside these
limits.
v) 99% values lie inside x    2.58 to x    2.58 , i.e. only 1% values lie outside
these limits.
# Here onwards we will denote the pdf of standard normal variable z by symbol
f ( z) .
# Values of pdf of standard normal variate can be obtained from statistical tables.
Illustrative Examples
Q 1) A certain number of articles manufactured one batch were classified into 3 categories
according to a particular characteristic, being less than 50, between 50 – 60 and  60 . If
 Random Variable
this characteristic is known to be normally distributed, determine the mean and s.d. for
this batch if 60%, 35% and 5% were found in these categories.
X 
Soln. Let Z  .

 60   
P( Z1 )  P  Z    0.6
  
 60   
P( Z  Z 2 )  P  Z    0.05
  
50  
A1  Z1   0.1

60  
A2  Z 2   0.05

6  7.543,   48.092
Z1 Z1 Z1



f ( z )dz  0.6  0.5   f ( z )dz  0.6
0


f ( z )dz  0.1  Z1  0.26 .

Q 2) In a normal distribution, 31% items are under 45 and 8% over 64. Find mean and s.d.

X  45  
Z1

Soln. Z
6
S.N.V. Z1 

P ( Z  Z1 )  

f ( z )dz
0 Z1 Z1

0.31  

f ( z )dz   f ( z)dz   f ( z )dz  0.19
0 0

64  
From table, Z1  0.5 Z2 

  Z2

P( Z  Z 2 )  
 Z2
f ( z )dz   f ( z )dz   f ( z )dz
0 0
0.5 Z2


0
f ( Z )dZ  0.19   f ( z )dz  0.5  0.08  0.42
0

45   64  
From table, Z 2  1.4  0.5 &
Thus,  1.4    50 &   10
 
Q 3) A manufacturer of air-mail envelopes knows from experience that the weight of envelope
is normally distributed with mean 1.95 gm and s.d. 0.05 gm. About how many envelopes
weighing 2 gm or more 2.05 gm or more can be expected in a given packet of 100
envelopes.
Soln.   1.95 &   0.05
 Random Variable
 
 2  1.95 
1
PZ    P( Z  1)   f ( z )dz   f ( z )dz   f ( z )dz  0.5  0.3413  0.1587
 0.05  1 0 0

 No. of envelopes weighing  2 gm  100  0.1587  15.87  16

 
 2.05  1.95 
2
PZ    P( Z  2)   f ( z )dz   f ( z )dz   f ( z )dz  0.5  0.4772  0.0228
 0.05  2 0 0

 No. of envelopes weighing >2.05  2.28  2

Q 4) The mean height of 500 students is 151 cm and the s.d. is 15 cm. Assuming that heights
are normally distributed, find how many students’ height lie between 120 and 155.
 120  151   155  151 
Soln. P Z    P(2.0667  Z  0.2667)
 15   15 
 P(2.07  Z  0.27)
0.27 0 0.27
 
2.07
f ( z )dz  
2.07
f ( z )dz  
0
f ( z )dz  0.4808  0.1064  0.5872

No. of students  296.3  296

 Joint Probability Distribution
The joint distribution of two random variables is called ‘Bivariate Probability Distribution’.
Bivariate Discrete Random Variables
Joint random variable ( X , Y ) is discrete if and only if X & Y both are discrete.
Let X and Y be two random variables taking values x1 , x2 , , xm and y1 , y2 , , yn respectively.
Then the range of bivariate variable ( X , Y ) is
R( X , Y )  R( x , y )  ( xi , y j ): 1  i  m,1  j  n , i. e.,
R ( X , Y )  R( x , y ) is the Cartesian product of ranges of X and Y . R( x , y )  Rx  Ry .
Let ( X , Y ) be a discrete random variable defined on sample space S .
R( x , y )  Range of X and Y . For each ( xi , y j ) define a function, P ( xi , y j )  Pi j .
Pi j  P [ X  xi , Y  y j ] . This gives a probability mass function of ( X , Y ) if following axioms
are satisfied.
Axioms
m n
1) 0  P ( xi , y j )  1 2)  P( x , y )  1
i 1 j 1
i j

3) P [( xi  A, y j  B)  ( xi  C , y j  D)]  P [( xi  A, y j  B)]  P [( xi  C , y j  D)]

Pmf of joint distribution can be tabulated as follows :
X
y1 y2 yn Total
Y
x1 P11 P12 P1n P( X  x1 )
x2 P21 P22 P2n P( X  x2 )

xm Pmn P( X  xm )
Total P(Y  y1 ) P(Y  y2 ) P(Y  yn ) 1
 Random Variable
Marginal Distributions
Let {( xi , y j ), P( xi , y j )), i  1, , m, j  1, , n} represents a joint probability
distribution of a bivariate random variable ( X , Y ) .
n
The marginal distribution of X is obtained as P[ X  xi ]   P [ X  xi , Y  y j ] .
j 1

This can be tabulated as

X x1 x2 ………………….. xm
P[ X  x j ] P[ X  x1 ] P[ X  x2 ] P[ X  xm ]
Total 1
Similarly, The marginal distribution of Y is obtained as
m
P(Y  y j )   P[ X  xi , Y  y j ] .
i 1
The marginal distribution for Y can be tabulated as
Y y1 y2 ………………….. yn
P[Y  x j ] P[Y  y1 ] P[Y  y2 ] P[Y  yn ] Total1
Covariance :
It is a joint variance between X and Y . It is given by
Cov( X , Y )  E ( XY )  E ( X ) E (Y ) , where
m n
E ( X )   xi P ( X  xi ) , E (Y )   y j P[Y  y j ] and E ( XY )   xi y j P[ X  xi , Y  y j ] .
n m

i 1 j 1 j 1 i 1

CorelationCoefficient :
It is a measure of correlation between the two random variates. It is given by the formula
Cov( X , Y ) Cov( X , Y )
  .
sd ( X ) sd (Y )  X Y
Note that :  always lie between 1 and 1 . Depend on the value of  , following
conclusions can be drawn about the variables
   0  Variatesare uncorrelated.
 0    1  Variates are negatively correlated.
 1    0  Variates are positively correlated.
   1  There is perfect positive correlation.
   1  There is perfect negative correlation.
Independent Random Variables:
Two random variables are said to be independent if the product of marginal probabilities
equal to the corresponding joint probability.
Pi j  P( X  xi )  P(Y  y j ) i, j .
Note that: Two Independent random variables are always Uncorrelated. But the converse
is not true. This means Uncorrelated Independence.
 Random Variable
Illustrative Example
Q 1) Consider the following bivariate distribution.
X
0 1 P( X ) For ( X )
Y
1 1
1 0
3 3
1 1
0 0
3 3
1 1
1 0
3 3
1 2
P(Y ) 1  For (Y )
3 3

1. Write the marginal distribution for X and Y .

2. Find Cov( X , Y ) .
3. Find  ( X , Y )
4. Write the conclusion about value of  ( X , Y ) .
5. Are the random variables Independent?
Soln. Marginal Distribution for X

X -1 0 1
P( X ) 1/3 1/3 1/3 1

Marginal Distribution for Y

Y 0 1
P(Y ) 1/3 2/3 1

1 1 1 1 2 2
E ( X )  1  0   1  0 and E (Y )  0   1 
3 3 3 3 3 3
1 1 1
E ( XY )  1 0  1  0   0  0  0  0  1  0
3 3 3
Cov( X , Y )  E ( XY )  E ( X ) E (Y )  0 .  ( X , Y )  0 .
Therefore variates are uncorrelated.
1 1
Now P11  0  P( X  1) P(Y  0)  
3 3
X and Y are not independent.
Q 2) Let X and Y be independent random variables with the following distributions :

X 1 2 Y 5 10 15
p( X  xi ) 0.6 0.4 p (Y  y j ) 0.2 0.5 0.3
Find the joint distribution of X and Y .
 Random Variable
Soln.Since X and Y are independent, joint probabilities are the product of marginal
probabilities. Hence the joint distribution of X and Y is
Y 5 10 15 Marginal Probabilities
X of X
1 0.12 0.30 0.18 0.6
2 0.08 0.20 0.12 0.4
Marginal
0.2 0.5 0.3 Total 1
Probabilities of Y

Q 3) A fair coin is tossed three times. Let X equals 0 or 1 according as a head or a tail occurs
on the first toss. Let Y equals the total number of heads that occur in each toss.
a) Write the marginal distributions of X and Y .
b) Write the joint distribution of X and Y .
c) Find Cov  X , Y 
d) Find   X , Y  . Comment on value of   X , Y  .
e) Determine whether X and Y are independent.
Soln. The sample space of the experiment is
HHH , HHT , HTH , THH , HTT , THT , TTH , TTT 
a) For r. v. X we have X  HHH   X  HHT   X  HTH   X  HTT   0
And X THH   X THT   X TTH   X TTT   1
For r. v. Y we have Y  HHH   3 , Y  HHT   Y  HTH   Y THH   2 ,
Y THT   Y TTH   Y  HTT  1 , Y TTT   0 .
Marginal Distributions
p  X  0   p HHH    HHT   HTH   HTT    .
1
2
p  X 1  p THH   THT   TTH   TTT   
1
2
Hence the marginal distribution of X
X 0 1
p( X  xi ) 1 1
2 2

p Y  0   p TTT    p Y  1  p  HTT   THT   TTH   

1 3
, ,
8 8
p Y  2   p  HHT   THH    HTH    p Y  3  p  HHH    .
3 1
,
8 8
Hence the marginal distribution of Y is

Y 0 1 2 3
p (Y  y j ) 1 3 3 1
8 8 8 8
 Random Variable

b) Joint Distribution X and Y

p  X  0, Y  0  p    0 , p  X  0, Y  1  p  HTT   
1
,
8
p  X  0, Y  2   p HHT   HTH    , p  X  0, Y  3  p  HHH    ,
1 1
4 8
p  X  1, Y  0   p TTT    , p  X  1, Y  1  p THT   TTH    ,
1 1
8 4
p  X  1, Y  2   p THH    , p  X  1, Y  3  p    0 .
1
8
Hence the joint distribution is

Y 0 1 2 3 Marginal
X Probabilitiesof X

0 0 1/ 8 1/ 4 1/ 8 1/ 2

1 1/ 8 1/ 4 1/ 8 0 1/ 2

Marginal 1/ 8 3/ 8 3/ 8 1/ 8 Total 1
Probabilities of Y

, E Y    y j p Y  y j   .
1 3
c) E  X    xi p  X  xi  
i 2 j 2

E  XY    xi y j p  X  xi , Y  y j    .
1
i, j 4
Cov  X , Y 
d)   X , Y   ,  X  Var  X  ,  Y  Var Y  .
 X . Y
Var  X   E  X 2    E  X   , E  X 2    xi 2 p  X  xi   . Var  X   .
2 1 1
Therefore
i 2 4

  
 X  . . Var Y   E Y 2    E Y   , E Y 2   y j 2 p Y  y j  3 .
1
2
2

j

3 3 Cov  X , Y   1/ 4    1 .
Var Y   . Therefore Y  .   X ,Y   
4 2  X . Y 
1/ 2  3 / 2 3
  X , Y   0, variates X and Y are negatively correlated.
1 1
e) p  X  0   , p Y  0   But p  X  0, Y  0  0  p  X  0 p Y  0 .
2 8
Hence variables are not independent.