29-Mar-13 Dr. B.

Mishra (Prob & Stat) 1

Chapter 4
Continuous Distributions
29-Mar-13 Dr. B. Mishra (Prob & Stat) 2
Continuous Random Variables
In many situations, we come across random
variables that take all values lying in a certain
interval of the x-axis.
(a) The life length X of a bulb is a
continuous random variable that can take
all non-negative real values.
Examples
29-Mar-13 Dr. B. Mishra (Prob & Stat) 3
(b) The time between two consecutive
arrivals in a queuing system is a random
variable that can take all non-negative
real values.
(c) The distance R of the point (where a
dart hits) (from the centre) is a
continuous random variable that can
take all values in the interval (0,a)
where a is the radius of the board.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 4
It is clear that in all such cases, the
probability that the random variable takes
any one particular value is meaningless (and
is zero). We only ask for the probability that
the random variable takes values in an
interval. For example, when you buy a bulb,
you ask the question? What are the chances
that it will work for at least 500 hours? (That
is we want the probability that the life length
X is > 500.)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 5
Definition
(Continuous Random Variable)
X is called a continuous random variable
if its range is an interval (or intervals) of
real numbers and the probability that it
assumes any one specific value is zero.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 6
If X is a continuous random variable, the
questions about the probability that X takes
values in an interval (a, b) are answered by
defining a probability density function.
Probability Density function (pdf)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 7
( )
( ) 0 for all a f x x >
( )
( ) 1 b f x dx

=
}
( ) ( )
( ) X .
b
a
c P a b f x dx s s =
}
A real function f (x) is called the
probability density function of X if
for any two real numbers a, b
Definition
Let X be a continuous random variable.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 8
Condition (a) is needed as probability is
always > 0.
Condition (b) says that the probability of
the certain event is 1.
Condition (c) says to get the probability that
X takes a value between a and b, integrate the
probability density function f (x) between
a and b.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 9
P(a s X s b) a b
f(x)
( )
.
b
a
f x dx
}
29-Mar-13 Dr. B. Mishra (Prob & Stat) 10
( ) ( ) ( )
1. X X 0
a
a
P a P a a f x dx = = s s = =
}
( ) ( )
( ) ( )
2. X X
X X
P a b P a b
P a b P a b
s s = < s
= s < = < <
Remarks

Note: it is immaterial whether we include or
exclude one or both the end points.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 11
( ) ( )
3. X P x x x f x x s s +A ~ A
Thus in words, the probability that X takes
values in the interval [ x, x + Ax ] is
approximately f (x) times the length of the
interval, assuming that it is a small interval.
This is proved using Mean value theorem (of
Integral Calculus).
29-Mar-13 Dr. B. Mishra (Prob & Stat) 12
Cumulative Distribution function
( ) ( ) ( )
X X
x
P x P x f t dt

s = < s =
}
We denote the above by F(x) and call it the
cumulative distribution function (cdf) of X.
If X is a continuous r.v. and if f (x) is its
density,
29-Mar-13 Dr. B. Mishra (Prob & Stat) 13
F(x) = P( X s x)
- x
f (t)
F (x) = Area under f (t) from t = - to t = x
t
y
( )
x
f t dt

=
}
29-Mar-13 Dr. B. Mishra (Prob & Stat) 14
( )
(1) 0 1 for all . F x x s s
( ) ( )
1 2 1 2
(2) x x F x F x < s
Properties of cdf
i.e. F(x) is a non-decreasing function of x.

29-Mar-13 Dr. B. Mishra (Prob & Stat) 15
( ) ( ) ( ) ( )
(3) 0; 1.
lim lim
x x
F F x F F x

= = + = =
( ) ( ) ( )
(4)
x
d d
F x f t dt f x
dx dx

= =
}
(Thus we can get density function f (x) by
differentiating the distribution function F (x)).
(5) ( ) P a X b s s = ( ) ( ) F b F a
29-Mar-13 Dr. B. Mishra (Prob & Stat) 16
Problem
If the probability density of a random variable
X is given by
2
0 1
( )
0
k x x
f x
elsewhere

< <
=

find the value of k and the probability that

the random variable takes on a value
(a) between and ; (b) greater than 2/3 .
29-Mar-13 Dr. B. Mishra (Prob & Stat) 17
Problem
With reference to the preceding exercise, find
the corresponding distribution function and
use it to determine the probabilities that a
random variable having this distribution
function will take on a value
(a) greater than 0.8; (b) between 0.2 and 0.4.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 18
Problem
If the probability density of a random variable
is given by
for 0 1
( ) 2 for 1 2
0 elsewhere
x x
f x x x
< <

= < <

find the probabilities that a random variable

having this probability density will take on a
value (a) between 0.2 and 0.8; (b) between 0.6
and 1.2.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 19
Problem
With reference to the preceding exercise, find
the corresponding distribution function and
use it to determine the probabilities that a
random variable having this distribution
function will take on a value
(a) greater than 1.8; (b) between 0.4 and 1.6.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 20
Exercise 4.1(14a)
Decide whether the following function
can be the cdf of a continuous rv X. If yes
find the corresponding pdf; if not explain
what property fails.
0 1
( ) 1 1 0
1 0
x
F x x x
x
<

= + s <

>

29-Mar-13 Dr. B. Mishra (Prob & Stat) 21

Exercise 4.1(14b)
Decide whether the following function can be
the cdf of a continuous rv X. If yes find the
corresponding pdf; if not explain what
property fails.
2
0 0
0 1/ 2
( )
/ 2 1/ 2 1
1 1
x
x x
F x
x x
x
s
< s
=
< s
>

29-Mar-13 Dr. B. Mishra (Prob & Stat) 22

Problem
The mileage (in thousands of miles) that car
owners get with a certain kind of tire is a
random variable having the probability
density
/ 20
1
for 0
( )
20
0 for 0
x
e x
f x
x

>

Expectation
and
Distribution Parameter
29-Mar-13 Dr. B. Mishra (Prob & Stat) 23
29-Mar-13 Dr. B. Mishra (Prob & Stat) 24
The density of a r.v. X is
( )
2
3 0 1
0 elsewhere
x x
f x

< <
=

Problem
Find its mean and variance.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 25
The density of a r.v. X is
( )
/ 20
1
0
20
0 elsewhere
x
e x
f x

>

Problem
Find its mean and variance.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 26
Moment Generating Function
Let X be a continuous r.v. with density f (x).
We define its mgf as
( )
X
X
( ) e
t tx
M t E e f x dx

(
= =

}
Note that the mgf exists (=is defined) only
for those t for which the RHS infinite
integral converges.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 27
Properties of MGF
1. For any r.v. X, M
X
(0) = 1.
2. If X is a r.v. with mgf M
X
(t) and if
Y = aX+b, then the mgf of Y is given by
M
Y
(t) = e
bt
M
X
(at)
3. If X and Y are independent random
variables, then
M
X+Y
(t) = M
X
(t) M
Y
(t)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 28
Some simple applications
1. The function
2
( )
1
g t
t
=

cannot be the mgf of any random variable

because
its value at t = 0 is 2 (and NOT 1).
2. If the mgf of X is
1
( )
1
X
M t
t
=

Then the mgf of Y = 3X + 2 is

2
1
( )
1 3
t
Y
M t e
t
=

29-Mar-13 Dr. B. Mishra (Prob & Stat) 29

Let X and Y be independent random
variables each having Poisson distribution
with parameters and respectively.
Hence mgf of X is M
X
(t) =
( 1)
t
e
e

Hence mgf of Y is M
Y
(t) =
( 1)
t
e
e

Hence mgf of X+Y is
( )( 1)
t
e
e
+
29-Mar-13 Dr. B. Mishra (Prob & Stat) 30
Hence we can say that sum of two
independent Poisson random variables is
also a Poisson random variable.
The parameter of the sum, X + Y, is the sum
of the parameters of X and Y.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 31
Exercise: 4.2(15)
Consider the random variable X with
density
1
( ) , 2 4
6
f x x x = s s
(a) Find E[X]
(b) Find E[X
2
]
(c) Find o
2
, and o.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 32
Let X denote the amount in pounds of
polyurethane cushioning found in a car.
The density for X is given by
1 1
( ) , 25 50
ln2
f x x
x
= s s
Find the mean, variance and standard
deviation of X.
Exercise: 4.2(16)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 33
Exercise: 4.2(17)
Let X denote the length in minutes of a long
distance telephone conversation. The density
of X is given by
10
1
( ) , 0
10
x
f x e x

= >
(a) Find the mgf of X.
(b) Using the mgf find the mean, variance
and standard deviation of X.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 34
Let X be a random variable with Cauchy
Distribution. The density for X is given by
2
1 1
( ) ,
1
f x x
x t
= < <
+
Show that E [X] does not exist.
Exercise: 4.2(22)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 35
Let X denote the amount of time in hours
that a battery on a solar calculator will
operate adequately between exposures to
light sufficient to recharge the battery.
Assume that the density for X is given by
3
50 1
( ) , 2 10
6
f x x
x
= < <
(a) Verify that this is a valid continuous density.
Exercise: 4.2(23)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 36
(b) Find the expression for the cdf for X and
use it to find the probability that a randomly
selected solar battery will last at most 4
hours before needing to be recharged.
(c) Find the average time that a battery will
last before needing to be recharged.
(d) Find E[X
2
] and use this to find the
variance of X.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 37
Assume that the increase in demand for
electric power in millions of kilowatt
hours over the next 2 years in a particular
area is a random variable whose density
is given by
3
1
( ) , 0 4
64
f x x x = < <
(a) Verify that this is a valid continuous
density.
Exercise: 4.2 (24)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 38
(b) Find the expression for the cdf for X and
use it to find the probability that the
increase in demand will be at most 2 million
kilowatt hours.
(c) If the area has only the capacity to
generate an additional 3 million kilowatt
hours, what is the probability that the
demand will exceed supply?
(d) Find the average increase in demand.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 39
Uniform Distribution
Gamma Distribution
Exponential Distribution
Chi-squared Distribution

29-Mar-13 Dr. B. Mishra (Prob & Stat) 40
The Uniform Distribution
A r.v X is said to have uniform distribution
over the interval (o, |) if its density is given
by
( )
1
0 elsewhere
x
f x
o |
| o

< <

Thus the graph of the density is a constant

over the interval (o, |) .
29-Mar-13 Dr. B. Mishra (Prob & Stat) 41
Graph of the uniform density
o
|
f (x)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 42
( )
P c X d < < =
If o < c < d < |,
and thus is proportional to the length
of the interval ( c, d ).
You may verify that the mean of X is
( )
2
| o

+
= = = X E X
( )
2
E X
o |

+
= =
(mid point of the interval
(o , |))
1
d
c
dx
| o
=

}
d c
| o

29-Mar-13 Dr. B. Mishra (Prob & Stat) 43

The variance of X is
( )
2
2
12
| o
o

=
The cumulative distribution function (cdf) is
( )
0
1
x
x
F x x
x
o
o
o |
| o
|
<

= s s

>

29-Mar-13 Dr. B. Mishra (Prob & Stat) 44

Proof:
2
o | +
=
( )
E X = =
1
x dx
|
o
| o
=

}
2
1
2
x
|
o
| o
(
=
(

2 2
1
2
| o
| o

=

( )
2
E X
o |

+
= =
( ) x f x dx

}
29-Mar-13 Dr. B. Mishra (Prob & Stat) 45
Proof:
2 2
3
| o| o + +
=
( )
2
E X =
2
1
x dx
|
o
| o
=

}
3
1
3
x
|
o
| o
(
=
(

3 3
1
3
| o
| o

=

( )
2
2
12
| o
o

=
2
( ) x f x dx

}
29-Mar-13 Dr. B. Mishra (Prob & Stat) 46
Hence the variance equals
( )
2 2 2
E X o =
2
2 2
3 2
| |o o o | + + +
| |
=
|
\ .
2 2 2 2
2
3 4
| |o o o o| | + + + +
=
2 2
2
12
| |o o +
=
( )
2
12
| o
=
29-Mar-13 Dr. B. Mishra (Prob & Stat) 47
The cdf of the uniform density
Case (ii) o < x s |. In this case f (t) = 0 for
Distribution function is
( ) ( )dt t f x F
x
}

=
Case (i) x s o. In this case f (t) = 0 for all t s x.

Hence F(x) = 0.
t s o and f (t) = 1/(| - o) for o s t s x. Hence
29-Mar-13 Dr. B. Mishra (Prob & Stat) 48
1
( ) .
x
x
F x d t
o
o
| o | o

= =

}
0 s x
( ) 0 = t f
Case (iii) x > |. In this case f(t) = 0 for t s o
or t > | and f (t) = 1/(| - o) for o < t < |.
Hence
1
( ) 0 0
x
F x d t d t d t
o |
o |
| o

= + +

} } }
= 1.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 49
Thus the cdf of the uniform density is
( )
0
1
x
x
f x x
x
o
o
o |
| o
|
<

= s s

>

29-Mar-13 Dr. B. Mishra (Prob & Stat) 50

MGF of the uniform density
Let X have the uniform density over the
interval (o, |). Hence the density of X is
( )
1
, f x x o |
| o
= < <

Hence the mgf of X is

29-Mar-13 Dr. B. Mishra (Prob & Stat) 51
M
X
(t) = E[e
tX
] =
( )
tx
e f x dx

}
1
tx
e dx
|
o
| o
=

}
( )
( )
1
t t
e e
t
| o
| o
=

(Note that M
X
(0) exists and = 1.)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 52
In certain experiments, the error X made in
determining the solubility of a substance is a
random variable having the uniform density
with o = - 0.025 and | = 0.025. What are the
probabilities that such an error will be
Problem
(b) between 0.012 and 0.012?
(a) between 0.010 and 0.015?
29-Mar-13 Dr. B. Mishra (Prob & Stat) 53
Exercise: 4.1(6)
If a pair of coils were placed around a
homing pigeon and a magnetic field was
applied that reverses the earths field, it is
thought that the bird would become
disoriented. Under these circumstances it is
just as likely to fly in one direction as in any
other. Let u denote the direction in radians of
the birds initial flight. u is uniformly
distributed over the interval [0, 2].
29-Mar-13 Dr. B. Mishra (Prob & Stat) 54
(a) Find the density for u.
u
Home
Pigeon
u = direction of the initial flight of a
homing pigeon measured in radians
29-Mar-13 Dr. B. Mishra (Prob & Stat) 55
(b) Sketch the graph of the density. The
uniform distribution is sometimes called the
rectangular distribution. Do you see why?
(c) Shade the area corresponding to the
probability that a bird will come within /4
radians of home, and find this area using
plane geometry.
(d) Find the probability that a bird will orient
within /4 radians of home by integrating
the density over the appropriate region(s)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 56
and compare your answer with that
obtained in part (c).
(e) If 10 birds are released independently and
at least seven orient within /4 radians of
home, would you suspect that perhaps the
coils are not disorienting the birds to the
extent expected? Explain, based on the
probability of this occurring.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 57
From experience, Mr. Harris has found that
the low bid on a construction job can be
regarded as a r.v. X having uniform density
Problem
where C is his own estimate of the cost of the
job. What percentage should Mr. Harris add to
his cost estimate when submitting bids to
( )
3 2
2
4 3
0 elsewhere
C
x C
f x
C

< <

29-Mar-13 Dr. B. Mishra (Prob & Stat) 58

maximize his expected profit?
29-Mar-13 Dr. B. Mishra (Prob & Stat) 59
Gamma Function
1
0
t x
e t dt

}
converges to a finite real number which we
denote by I(x) (Capital gamma of x).
( ) x I
( ) x I
If x > 0, it is shown that the improper integral
29-Mar-13 Dr. B. Mishra (Prob & Stat) 60
Thus for all real no x > 0, we define
1
0
( )
t x
x e t dt

I =
}
Properties of Gamma Function
( ) ( )
1. 1 x x x I + = I
( )
2. 1 1 I =
( ) ( ) ( ) ( )
3. 2 1 1 1, 3 2 2 2 1 2! I = I = I = I = =
29-Mar-13 Dr. B. Mishra (Prob & Stat) 61
( )
1 ! for all +ve integers . n n n I + =
More generally
1
4.
2
| |
I
|
\ .
. t =
decreases in the interval (0,1)
( ) , 2
a minimum somewhere between 1 and 2.
( )
5. x I
and increases in the interval
and has
Graph of the Gamma function in the next slide.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 62
29-Mar-13 Dr. B. Mishra (Prob & Stat) 63
The Gamma Distribution
Let o, | be positive real numbers. A r.v. X is
said to have a Gamma Distribution with
parameters o > 0, | > 0 if its density is
( )
( )
.
1
1
0
0 elsewhere
x
e x x
f x
|
o
o
o |

>

I
=

29-Mar-13 Dr. B. Mishra (Prob & Stat) 64

It can be shown that
( )
Mean of X E X o| = = =
2 2
Variance of X . o o| = =
1
MGF of X (1 ) , t t
o
|
|

= <
29-Mar-13 Dr. B. Mishra (Prob & Stat) 65
Proof : Mean = o |
( )
Mean of X E X =
1 /
0
1
( )
x
x x e dx
o |
o
o |

=
I
}
Put
/ x u | =
0
( )
u
e u du
o
|
o

=
I
}
( 1)
( )
|
o
o
= I +
I
o| =
as I(o+1) = o I(o).
29-Mar-13 Dr. B. Mishra (Prob & Stat) 66
Proof: Variance = o |
2

( )
2
E X
2 1 /
0
1
( )
x
x x e dx
o |
o
o |

=
I
}
Put
/ x u | =
2
1
0
( )
u
e u du
o
|
o

+
=
I
}
2
( 2)
( )
|
o
o
= I +
I
2
( 1) o o| = +
Hence Variance = E[X
2
]-(E[X])
2
= o |
2
.
2 2 2
o | o| = +
as I(o+2) = (o +1)o I(o).
2 2
( [ ]) E X o| = +
29-Mar-13 Dr. B. Mishra (Prob & Stat) 67
Derivation of the MGF:
( )
X
(t)=E
t
X
M e
1 /
0
1
( )
tx x
e x e dx
o |
o
o |

=
I
}
Put
/ x u | =
(1 ) 1
0
1
( )
u t
e u du
| o
o


=
I
}
Put
( )
1 u t v | =
1
0
1
( )(1 )
v
e v dv
t
o
o
o |


=
I
}
Assume t < 1/|
1
(1 ) , ( ) t t
o
|
|

= <
29-Mar-13 Dr. B. Mishra (Prob & Stat) 68
Example
In a certain city, the daily consumption of
electric power (in millions of kilowatt-hours)
can be treated as a random variable X having
a Gamma distribution with parameters o = 3
and | = 2. If the power plant of this city has a
daily capacity of 12 million kilowatt-hours,
what is the probability that the power supply
will be inadequate on any given day?
29-Mar-13 Dr. B. Mishra (Prob & Stat) 69
Problem
If n salespeople are employed in a door-to-
door selling campaign, the gross sales
volume in thousands of dollars may be
regarded as a random variable having the
Gamma distribution with o = 100 and
| = . If the sales costs are $5,000 per
salesperson, how many salespeople should be
employed to maximize the expected profit?
n
29-Mar-13 Dr. B. Mishra (Prob & Stat) 70
The Exponential Distribution
If, in the gamma distribution, o = 1, we say X
has exponential distribution. Thus X has an
exponential distribution (with parameter | > 0)
if its density is
( )
/
1
0
0 elsewhere
x
e x
f x
|
|

>

29-Mar-13 Dr. B. Mishra (Prob & Stat) 71

Graph of the Exponential Density
0
1/|
f (x)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 72
We also see that
( )
1. Mean of X E X | = = =
2 2
2. Variance of X . o | = =
1
3. MGF of X , 1/
1
t
t
|
|
= <

29-Mar-13 Dr. B. Mishra (Prob & Stat) 73

4. The cumulative distribution function of X is
( )
/
1 0
0 elsewhere
x
e x
F x
|

>
=

29-Mar-13 Dr. B. Mishra (Prob & Stat) 74

4. X has the memoryless property:
( ) ( )
X | X X , , 0 P s t s P t s t > + > = > >
Proof of (4):
( ) ( )
X 1 X P s P s > = s
( )
/
1
s
F s e
|
= =
by (3).
29-Mar-13 Dr. B. Mishra (Prob & Stat) 75
( )
X | X P s t s > + >
Now
( ) ( )
( )
X X
X
P s t s
P s
( > + >

=
>
( )
( )
X
X
P s t
P s
> +
=
>
( ) /
/
s t
s
e
e
|
|
+

=
/ t
e
|
= ( )
P X t = >
29-Mar-13 Dr. B. Mishra (Prob & Stat) 76
Example
The amount of time that a surveillance
camera will run without having to be reset is
a random variable X having exponential
distribution with | = 50 days. Find the
probability that such a camera will
(a) have to be reset in less than 20 days;
(b) will not have to be reset in at least 60 days.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 77
Problem
Given a Poisson process with on the average
o arrivals per unit time, find the probability
that there will be no arrivals during a time
interval of length t, namely, the probability
that the waiting time between successive
arrivals will be at least of length t.
Let X
t
= Number of arrivals during a time
interval of length t
Solution
29-Mar-13 Dr. B. Mishra (Prob & Stat) 78
We know that X
t
is a random variable having
Poisson distribution with parameter o t.
Thus, the probability that there will be no
arrivals during a time interval of length t
( )
X 0
t
P = =
t
e
o
=
29-Mar-13 Dr. B. Mishra (Prob & Stat) 79
Given that the switchboard of a consultants
office receives on an average 0.6 calls per
minute (in a Poisson process), find the
probabilities that the time between
successive calls arriving at the switchboard
of the consultants office will be
Problem
(a) less than minute;
(b) more than 3 minutes.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 80
The occurrences of blemishes on a
continuously produced bolt of cloth follows
a Poisson process with an average of 0.01
per meter. Find the probabilities that the
length of cloth between two consecutive
blemishes is
(a) less than 10 meters.
(b) more than 50 meters.
Problem
29-Mar-13 Dr. B. Mishra (Prob & Stat) 81
Exercise: 4.3 (34)
A particular nuclear plant releases a
detectable amount of radioactive gases
twice a month on the average. Find the
probability that at least 3 months will elapse
before the release of the first detectable
emission. What is the average time one
must wait to observe the first emission?
29-Mar-13 Dr. B. Mishra (Prob & Stat) 82
Problem
Suppose that telephone calls arrive at a
switchboard in a Poisson fashion with an
average rate of 5 calls per minute. What is
the probability that up to a minute will elapse
until 2 calls have come in to the switchboard?
Solution
Unit of measurement is minute.
Average number of arrivals per unit of
measurement (= minute) = o = 5.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 83
Hence the time T until 2 arrivals is a random
variable having a Gamma distribution with
parameters 2 and 1/o i.e. 2 and 1/5.
Hence the density of T is
( )
( )
5 2 1
2
1
, 0
2 (1/ 5)
t
f t e t t

= >
I
( )
5
. . 25 , 0
t
i e f t te t

= >
29-Mar-13 Dr. B. Mishra (Prob & Stat) 84
Chi-Square Distribution
Definition
Let X be a random variable having Gamma
Distribution with parameters o = v/2, | = 2,
where v is a positive integer. We then say X is
a random variable having _
2
distribution with
v degrees of freedom.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 85
The density of a _
2
random variable with v
degrees of freedom is given by
( )
1
2 2
2
1
0
2
2
0 elsewhere
x
e x x
f x
v
v
v

>

| |
= I

|
\ .

29-Mar-13 Dr. B. Mishra (Prob & Stat) 86

It can be shown that
( )
Mean of X E X v = = =
2
Variance of X 2 . o v = =
2
1 1
MGF of X ,
2
(1 2 )
t
t
v
= <

29-Mar-13 Dr. B. Mishra (Prob & Stat) 87

o
_
o
2
P (_
2
> _
o
2
) = o
Graph of _
2
Density
29-Mar-13 Dr. B. Mishra (Prob & Stat) 88
If X = _
2
is a random variable having _
2

distribution with v degrees of freedom, then
we define _
o
2
(= ) to be that unique
number such that
2
, v o
_
P (_
2
> _
o
2
) = o
That is _
o
2
is that unique point on the
horizontal axis so that the are under the
density curve to the right of it is o. We
also sometimes write it as to indicate
the degrees of freedom.
2
,

29-Mar-13 Dr. B. Mishra (Prob & Stat) 89

In your book, _
1-o
2
has been tabulated for
various degrees of freedom v and for
various o in Table IV, Appendix A.
For example
_
0.95
2
(where v = 12) is 5.23
The intersection of row v = 12 and (4
th
)
column where 1 0.95 = 0.05
_
0.05
2
(where v = 18) is 28.9
29-Mar-13 Dr. B. Mishra (Prob & Stat) 90
Note that for v > 30, _
o
2
is to be approximated
by n + z
o


.
Hence for v = 50,
_
0.05
2

~ 50 +10 z
0.05
= 66.45
2n
Hence for v = 60,
_
0.01
2

~ 60 + z
0.01

120
= 60 + 10.9542.33
= 85.523
29-Mar-13 Dr. B. Mishra (Prob & Stat) 91
Exercise: 4.3(38)
Consider a chi-squared random variable
with 15 degrees of freedom.
(a) What is the mean of What is
its variance?
2
15
? _
Answer: Mean = v = 15
Variance = 2v = 30
29-Mar-13 Dr. B. Mishra (Prob & Stat) 92
(b) What is the expression for the density for
2
15
? _
(c) What is the expression for the mgf for
2
15
? _
29-Mar-13 Dr. B. Mishra (Prob & Stat) 93
(d) Use Table IV of Appendix A to find
each of the following:
P( s 5.23)
2
15
_
Answer
0.010
P( > 30.6)
2
15
_ 0.010
P(6.26 s s 27.5)
2
15
_
0.975 0.025 = 0.950
2 2
0.05 15,0.05
_ _ = 25.0
2 2
0.95 15,0.95
_ _ =
7.26
29-Mar-13 Dr. B. Mishra (Prob & Stat) 94
Normal Distribution
A random variable X is said to have a normal
distribution (also called Gaussian
distribution) with parameters and o (a
positive number) if its probability density is
given by
2
2
( )
2 2
1
( ; , ) ,
2
x
f x e x

o
o
t o

= < <
29-Mar-13 Dr. B. Mishra (Prob & Stat) 95
Normal Density Curve with mean
10 and s.d. 2
0
0.05
0.1
0.15
0.2
0.25
0 5 10 15 20
x
D
e
n
s
i
t
y
29-Mar-13 Dr. B. Mishra (Prob & Stat) 96
Properties of the Normal Distribution
The total area under the curve is equal to 1
The mean is and the variance is o
2

The distribution is bell shaped and
symmetrical about its mean
The curve extends to infinity in both
directions
The distribution is centered at the mean
It has two points of inflection, namely
x = - o and x = + o .
29-Mar-13 Dr. B. Mishra (Prob & Stat) 97
Proof : Mean
Put
2
x
t

=
2
1
2
1
( )
2
x
E X x e dx

o
t o

| |

|
\ .

=
}
2
1
Hence (X) ( 2 ) 2
2
t
E t e dt o o
t o

= +
}
2 2
2
0
t t
e dt t e dt
o

t t



= + = + =
} }
29-Mar-13 Dr. B. Mishra (Prob & Stat) 98
Proof: variance

2
(X ) E
Put
2
x
t

=
2
1
2 2
1
( )
2
x
x e dx

t o

| |

|
\ .

=
}
29-Mar-13 Dr. B. Mishra (Prob & Stat) 99
Hence

2
(X ) E
2 2
2
t t
t e e dt
o
t

(
=
`


)
}
{ }
2
2
0
o
t o
t
= + =
2
2 2
1
2 2
2
t
t e dt o o
t o

=
}
29-Mar-13 Dr. B. Mishra (Prob & Stat) 100
MGF of the Normal Distribution
( )
X
X
( ) e
t tx
M t E e f x dx

(
= =

}
2
1
2
1
2
x
t x
e e dx

o
t o

| |

|
\ .

=
}
Put
2
x
y

=
29-Mar-13 Dr. B. Mishra (Prob & Stat) 101
2
( 2 )
1
2
2
t y y
e e dy
o
o
t o

=
}
Now
2
( 2 )
1
t y ty
e e dy
o
t

=
}
2 2 2 2
1 1
( 2 ) ( )
2
2
y ty y t t o o o =
2
2 2
1
1
( )
2
2
1
y t
t t
e e dy
o
o
t

=
}
29-Mar-13 Dr. B. Mishra (Prob & Stat) 102
X
( ) M t
2
2 2
1
1
( )
2
2
1
y t
t t
e e dy
o
o
t


+

=
}
2 2
1
2
1
t t
e
o
t
t
+
=
2 2
1
2
X
( ) (all )
t t
M t e t
o +
=
Hence
Thus
29-Mar-13 Dr. B. Mishra (Prob & Stat) 103
A random variable having normal
distribution with mean = 0 and s.d.
o = 1 is said to have a standard normal
distribution and usually is denoted by
the symbol Z.
Standard Normal Distribution
29-Mar-13 Dr. B. Mishra (Prob & Stat) 104
Standard Normal curve
29-Mar-13 Dr. B. Mishra (Prob & Stat) 105
Let Z have standard normal distribution.
Then its cdf,
2
/ 2
1
2
z
t
e dt
t

=
}
cdf of the standard normal distribution
is tabulated for various values of z in Table
V.
F(z) = P(Z s z)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 106
For example, F(1.96) = P( Z s 1.96) = 0.9750
F(2.33) =
0.01
Can you relate F(z) and F(-z) ?
Answer: F(-z) = 1 F(z).
Also P ( a s Z s b) = F(b) F(a)
0.95
0.025
0.99 F(1.645) =
F(-1.96) = F(-2.33) =
Reason: By symmetry of the standard
normal curve.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 107
Thus all probability questions about Z can be
answered using F (z).
Now we ask:
What is that z for which F(z) = a given
value?
For example find z such that F(z) = 0.8438.
Answer:
z = 1.1
In particular
P ( - a s Z s a) = F(a) F(-a)
= 2F(a) 1
29-Mar-13 Dr. B. Mishra (Prob & Stat) 108
Notation: Given o, we denote by z
o
that
unique z such that P( Z > z
o
) = o. (That is
for which P( Z s z
o
) = 1- o.)
Thus
It is important to note that z
1-o
= -

z
o
We can now relate the probabilities that a
normal random variable X will have in terms
of F(z).
z
0.05
= 1.645,
z
0.01
= 2.33, z
0.95
= - 1.645
29-Mar-13 Dr. B. Mishra (Prob & Stat) 109
We note that if X is a random variable
having a normal distribution with mean
and s.d. o , then the random variable

has a standard normal distribution.
X
Z

o

=
Z is called the random variable obtained by
standardizing X.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 110
Thus P ( a s X s b)
.
b a
F F

o o

| | | |
=
| |
\ . \ .
Z
a b
P

o o

| |
= s s
|
\ .
29-Mar-13 Dr. B. Mishra (Prob & Stat) 111
Given a random variable having the
normal distribution with mean = 16.2
and variance o
2
= 1.5625, find the
probabilities that it will take on a value
(c) between 13.6 and 18.8;
Problem
(b) less than 14.9;
(d) between 16.5 and 16.7.
(a) greater than 16.8;
Exercise: 4.4 (41)
Most galaxies take the form of a flattened disc, with
the major part of the light coming from this very thin
fundamental plane. The degree of flattening differs
from galaxy to galaxy. In the milky way Galaxy
most gases are concentrated near the centre of the
fundamental plane. X is normally distributed with
mean 0 and standard deviation 100 parsecs ( A
parsec is equal to approximately 19.2 trillion miles).

112 Dr. B. Mishra (Prob & Stat) 29-Mar-13
(a) Sketch a graph of the density for X. Indicate on this
graph the probability that a gaseous mass is located
within 200 parsecs of the centre of the fundamental
plane. Find this probability.
(b) Approximately, what percentage of the gaseous
masses are located more than 250 parsecs from the
centre of the plane?
(c) What distance has the property that 20% of the
gaseous masses are at least this far from the
fundamental plane?
(d) What is the moment generating function for X.
-100 -50 50 100
0.0001
0.0002
0.0003
0.0004
0.0005
0.0006
0.0007
113 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Exercise: 4.4(43)
Let X denote the time in hours needed to locate
and correct a problem in the software that governs
the timing of traffic lights in the downtown area of
a large city. Assume X is normally distributed
with mean 10 hours and variance 9.
(a) Find the probability that the next problem will
requir at most 15 hours to find and correct.
(b) The fastest 5% of repairs take at most how
many hours to complete?
114 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Exercise: 4.4(44)
Assuming that during seasons of normal rainfall
the water level in feet at a particular lake follows
a normal distribution with mean of 1876 feet and
standard deviation of 6 inches.
(a) During such a season, would it be unusual to
observe a water level of at most 1875 feet.
(b) Suppose that the water will crest the spillway
if the water level exceeds 1878 feet. What is the
probability that this will occur during a season of
normal rainfall?
115 Dr. B. Mishra (Prob & Stat) 29-Mar-13
29-Mar-13 Dr. B. Mishra (Prob & Stat) 116
Exercise: 4.5 (48)
The number of Btus of petroleum and
petroleum products used per person in the US
in 1975 was normally distributed with mean
153 million Btus and standard deviation of 25
million Btus. Approximately, what percentage
of the population used between 128 and 178
million Btus during that year? Approximately
what percentage of the population used in
excess of 228 million Btus?
Log-Normal Distribution
The log-normal distribution is the
distribution of a random variable whose
natural logarithm follows a normal
distribution. Thus if X is a normal random
variable, then Y=e
X
follows a log-normal
distribution.
117 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Exercise: 4.4 (45)
Let X be normal with mean and variance o
2
. Let
G denote the cumulative distribution function for
Y = e
X
and let F denote the cumulative
distribution for X.
a) Show that G(y) = F(lny)
b) Show that G(y) =F(lny)/y
c) Show that the density for Y is given by



( )
. 0 , ); , ( ,
2
1
) (
2
2
2
ln
> e =

o
o t
o

y e
y
y g
y
118 Dr. B. Mishra (Prob & Stat) 29-Mar-13

Now, G(y) = P(Ysy) = P(e
X
s y) = P(X s lny)


( )
. 0 ); , ( , ,
2
1
) ln ( ) (
ln
2
2
2
> e = s =
}

o
o t
o

x dx e y X P y G
y
x
( )
x
y
y
e y where
y y d e y G or
=
> e =
}

. 0 , ); , ( ), (ln
2
1
) ( ,
0
2
ln
2
2
o
o t
o

( )
2
2
ln
2
0
1
, ( ) (ln ) (ln )
2
y
y
or G y e d y F y

o
t o

= =
}
119 Dr. B. Mishra (Prob & Stat) 29-Mar-13

Hence, show that the density for Y is given





g(y) = G(y)


( )
. 0 , ); , ( ,
2
1
) (
2
2
2
ln
> e =

o
o t
o

y e
y
y g
y
( )
) (ln ) (ln
2
1
) (
2
2
2
ln
y F y d e y G
y
y
= =
}

o

o t
120 Dr. B. Mishra (Prob & Stat) 29-Mar-13
29-Mar-13 Dr. B. Mishra (Prob & Stat) 121
In a photographic process, the developing time
of prints may be looked upon as a random
variable having normal distribution with mean
16.28 seconds and s.d. 0.12 second. Find the
probability that it will take
Problem
(a) anywhere from 16.00 to 16.50 seconds to
develop one of the prints;
(b) at least 16.20 seconds to develop one of
the prints;
29-Mar-13 Dr. B. Mishra (Prob & Stat) 122
(c) at most 16.35 seconds to develop one of
the prints.
29-Mar-13 Dr. B. Mishra (Prob & Stat) 123
With reference to the preceding exercise, for
which value is the probability 0.95 that it will
be exceeded by the time it takes to develop
one of the prints?
i.e. Find c such that P ( X > c) = 0.95
Problem
Normal Approximation to the Binomial
Distribution
Let X be binomial with parameters
n and p. For large n, X is
approximately normal with mean
np and variance np(1-p);
i.e.
( ) ) 1 ( , ~ p np np N X
124 Dr. B. Mishra (Prob & Stat) 29-Mar-13
For most practical purposes the
approximations is acceptable for
values of n and p such that either
p.5 and n.p>5
or p>.5 and n.(1-p)>5

125 Dr. B. Mishra (Prob & Stat) 29-Mar-13
126 Dr. B. Mishra (Prob & Stat) 29-Mar-13
127 Dr. B. Mishra (Prob & Stat) 29-Mar-13
128 Dr. B. Mishra (Prob & Stat) 29-Mar-13

Correction for continuity:

1.P(a < X b)=P(a-0.5 < X b+0.5)
2. P(X b)= P(- < X b) = P(- < X b+0.5)
= P(X b+0.5)
3. P(X a)= P(a X < )= P(a-0.5 X < )
=P( X a-.5)
4. P(X = a)=P(a-0.5 < X < a+0.5)
The number 0.5 is called the half-unit
correction factor for continuity.

129 Dr. B. Mishra (Prob & Stat) 29-Mar-13
29-Mar-13 Dr. B. Mishra (Prob & Stat) 130
A manufacturer knows that, on average, 2%
of the electric toasters that he makes will
require repairs within 90 days after they are
sold. Use normal approximation to the
binomial distribution to determine the
probability that among 1200 of these toasters
at least 30 will require repairs within the first
90 days after they are sold?
Problem
29-Mar-13 Dr. B. Mishra (Prob & Stat) 131
A safety engineer feels that 30% of all
industrial accidents in her plant are caused by
failure of employees to follow instructions.
Find approximately the probability that
among 84 industrial accidents in this plant
anywhere from 20 to 30 (inclusive) will be
due to failure of employees to follow
instructions.
Problem
(Normal Approximation to the Poisson Distribution)
Let X be Poisson with parameter s.Then
for large values of s, X is approximately
normal with mean s and variance s.






,....... 3 , 2 , 1 , 0 ,
!
) ( =

=
|
|
|
|
.
|

\
|
x
x
x
s
s
e
x f

132 Dr. B. Mishra (Prob & Stat) 29-Mar-13

133 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Exercise:4.6 (56)
Let X be a poisson random variable with
parameter s = 15. Find P[X 12] from the
table. Approximate this probability using a
normal curve. Be sure to employ the half-
unit correction factor.
134 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Exercise:4.6 (57)
The average number of jets either
arriving at or departing from OHare
Airport is one every 40 seconds. What
is the approximate probability that at
least 75 such flights will occur during a
randomly select hour? What is the
probability that fewer than 100 flights
will take place in an hour?
135 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Normal Probability Rule
Let X be normally distributed with parameters
mean and standard deviation o .Then
P[-o < X- < o ]=0.68
P[-2o < X- < 2o ]=0.95
P[-3o < X- < 3o ]=0.997
136 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Let X be a r.v. having mean and s.d. o. Then
for any positive number k, the probability of
getting a value of X which deviates from by
at least k o is at most 1/k
2
.
In symbols
( )
2
1
| | P X k
k
o > s
Chebyshevs Inequality
29-Mar-13 137 Dr. B. Mishra (Prob & Stat)
Proof
-ko
+ko
R
1

R
2

R
3

29-Mar-13 138 Dr. B. Mishra (Prob & Stat)
We divide the range of X into three regions
R
1
, where X s -ko
R
2
, where -ko < X < +ko
R
3
, where X > +ko
We assume that X is a continuous random
variable with density f (x).
29-Mar-13 139 Dr. B. Mishra (Prob & Stat)
Variance of X =
2 2
( ) ( ) x f x dx o

=
}
2 2
2
( ) ( ) ( ) ( )
( ) ( )
k k
k
k
x f x dx x f x dx
x f x dx
o o
o
o

+
= +
+
} }
}
29-Mar-13 140 Dr. B. Mishra (Prob & Stat)
Neglecting the middle term which is > 0, we get
In the region R
1
,i.e. (-, - ko )
x s -ko or x- s -ko and so (x - )
2
> k
2
o
2

In the region R
3
, i.e. (-, + ko ),
x > +ko or x- > ko and so (x - )
2
> k
2
o
2

2 2 2
( ) ( ) ( ) ( )
k
k
x f x dx x f x dx
o
o
o

+
> +
} }
29-Mar-13 141 Dr. B. Mishra (Prob & Stat)
Hence
Or
Now
and
29-Mar-13 142 Dr. B. Mishra (Prob & Stat)
2 2 2 2 2
( ) ( )
k
k
k f x dx k f x dx
o
o
o o o

+
> +
} }
2
1 ( ) ( )
k
k
k f x dx f x dx
o
o

+
| |
> +
|
|
\ .
} }
( ) ( )
k
P X k f x dx
o
o

s =
}
( ) ( )
k
P X k f x dx
o
o

+
> + =
}
Hence we get
Or
29-Mar-13 143 Dr. B. Mishra (Prob & Stat)
( )
2
1 ( ) ( ) k P X k P X k o o > s + >
2
1
( ) P X k
k
o > s
( )
2
k P X k o > >
29-Mar-13 Dr. B. Mishra (Prob & Stat) 144
If X is a discrete random variable, the
above proof holds by replacing integral by
summation:

i.e. replace by etc.
k o

}
k o

The event { |X - | < ko } is complementary

to the event { |X - | > ko }. Hence we can
also say that
Putting ko = c, we can also say that
( )
2
1
| | 1 P X k
k
o < >
( )
2
2 2
Var
| |
X
P X
o
c
c c
> s =
29-Mar-13 145 Dr. B. Mishra (Prob & Stat)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 146
Normal Distribution and Chebyshevs
inequality
Let X be a random variable with mean
and s.d. o. Then by Chebyshevs equality,
( 3 ) P X o < >
1
1 0.8999
9
=
But from normal tables, we find
( 3 ) P X o < =
(| | 3) P Z < =
2 (3) 1 F =
0.9974
29-Mar-13 Dr. B. Mishra (Prob & Stat) 147
Then by Chebyshevs inequality,
( 2 ) P X o < >
1
1 0.75
4
=
But from normal tables, we find
( 2 ) P X o < =
(| | 2) P Z < =
0.9544
( ) P X o < >
1
1 0.00
1
=
But from normal tables, we find
( ) P X o < =
(| | 1) P Z < =
0.6826
By Chebyshevs inequality,
In 1 out of 6 cases, material for bulletproof
vests fail to meet puncture standards. If 405
specimens are tested, what does
Chebyshevs inequality tell us about the
probability of getting at most 30 or at least
105 cases that do not meet puncture
standards?
Problem
29-Mar-13 148 Dr. B. Mishra (Prob & Stat)
Show that for 1 million flips of a balanced
coin, the probability is at least 0.99 that the
proportion of heads will fall between 0.495
and 0.505.
Problem
29-Mar-13 149 Dr. B. Mishra (Prob & Stat)
How many times do we have to flip a
balanced coin to be able to assert with a
probability of at most 0.01 that the difference
between the proportion of tails and 0.50 will
be at least 0.04?
Problem
29-Mar-13 150 Dr. B. Mishra (Prob & Stat)
29-Mar-13 Dr. B. Mishra (Prob & Stat) 151
Weibull Distribution
A r.v X is said to have a Weibull distribution
with parameters o >0, | > 0, and if its
density is given by
( )
1 ( )
( ) ,
x
f x x e x
|
| o
o|

= >
29-Mar-13 Dr. B. Mishra (Prob & Stat) 152
It can be shown that
( )
1/
Mean of X E X (1 1/ )
|
o |

= = = + I +
2
Variance of X o =
( )
2
2/
(1 2/ ) (1 1/ )
|
o | |

(
= I + I +

29-Mar-13 Dr. B. Mishra (Prob & Stat) 153
Proof: Mean =
( )
Mean of X E X =
1 ( )
( )
x
x x e dx
|
| o

o|

=
}
Put
( ) x u
|
o =
1/
0
u
u
e du
|

(
| |
= +
(
|
\ .
(

}
1/
(1 1/ )
|
o |

+ I +
1/
(1 1/ )
|
o |

= + I +
29-Mar-13 Dr. B. Mishra (Prob & Stat) 154
Variance:
( )
2
E X
( )
1
2 ( ) x
x x e dx
|
|
o

o|

=
}
Put
( 1) x u
|
o =
Hence Variance
2/ 1/
2
0
2
u
u u
e du
| |

o o

(
| | | |
= + +
(
| |
\ . \ .
(

}
2/ 1/ 2
(1 2/ ) 2 (1 1/ )
| |
o | o |

= I + + I + +
( )
2
2/
(1 2/ ) (1 1/ )
|
o | |

(
= I + I +

Transformation of Variables
Let X be a continuous random variable with
density . Let , where is strictly
monotonic and differentiable. The density for
Y is denoted by and is given by
155 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Monotonic functions
Monotonic functions are functions that tend to
move in only one direction as x increases.
A monotonic increasing function always increases
as x increases i.e. f(a) > f(b), for all a > b.
A monotonic decreasing function always
decreases as x increases i.e. f(a)<f(b), for all a>b.
In calculus speak,
A monotonic decreasing functions derivative is
always negative.
A monotonic increasing functions derivative is
always positive.

156 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Examples
(1) Let X be a random variable with density

Let . Find the density for
Y.
(2) Let X be uniformly distributed over (0,4) and
let .Find the density for
Y.

157 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Problem-70
Let X be a random variable with density


Let .
(a) Find E[X] and use the rules for
expectation to find E[Y].
(b) Find the density for Y.
(c) Use the density for Y to find E[Y].
158 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Problem-71
Let X be a random variable with density


Let . Find the density for Y.
159 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Problem-72
Let X be a random variable with density


Let . Find the density for Y.
160 Dr. B. Mishra (Prob & Stat) 29-Mar-13
Problem-74
Let X denote the velocity of a random gas
molecule. According to the Maxwell-Boltzmann
law, the density for X is given by


Here c is a constant that depends on the gas
involved, and is a constant whose value depends
on the mass of the molecule and its absolute
temperature. The kinetic energy of the molecule,
Y, is given by Y = (1/2)mX
2
, where m > 0. Find
the density for Y.
161 Dr. B. Mishra (Prob & Stat) 29-Mar-13
29-Mar-13 Dr. B. Mishra (Prob & Stat) 162
Exercise:4.8(75)
Let X be a continuous random variable with
density f
X
(x). Let Y = X
2
. Then the density
of Y is given by
{ }
1
( ) ( ) ( ) , 0
2
Y X X
f y f y f y y
y
= + >
29-Mar-13 Dr. B. Mishra (Prob & Stat) 163
Proof
Let y > 0.
( X ) y y = s s
Then the cdf of Y = G(y) = P( Y s y)
(F (x) is the cdf of X).
( ) ( ) F y F y =
Hence the density of Y = g (y)
( )
d
G y
dy
= { }
1
( ) ( ) , 0
2
X X
f y f y y
y
= + >
29-Mar-13 Dr. B. Mishra (Prob & Stat) 164
Exercise:4.8(76d)
Let Z have standard normal distribution. Then
the random variable Y = Z
2
has a chi-square
distribution with one degree of freedom.
Problem-77
Let X be normally distributed with mean
and variance
2
. Let Y = e
X
. Show that Y
follows the log-normal distribution.

165 Dr. B. Mishra (Prob & Stat) 29-Mar-13
29-Mar-13 Dr. B. Mishra (Prob & Stat) 166
Exercise:4.8(78)
Let Z be a standard normal variable and let
Y = 2Z
2
1. Find the density of Y.