Bivariate Continuous

UNIT 7 BIVARIATE CONTINUOUS Random Variables

RANDOM VARIABLES
Structure
7.1 Introduction
Objectives
7.2 Bivariate Continuous Random Variables
7.3 Joint and Marginal Distribution and Density Functions
7.4 Conditional Distribution and Density Functions
7.5 Stochastic Independence of Two Continuous Random Variables
7.6 Problems on Two-Dimensional Continuous Random Variables
7.7 Summary
7.8 Solutions/Answers

7.1 INTRODUCTION
In Unit 6, we have defined the bivariate discrete random variable (X, Y), where
X and Y both are discrete random variables. It may also happen that one of the
random variables is discrete and the other is continuous. However, in most
applications we deal only with the cases where either both random variables are
discrete or both are continuous. The cases where both random variables are
discrete have already been discussed in Unit 6. Here, in this unit, we are going
to discuss the cases where both random variables are continuous.
In Unit 6, you have studied the joint, marginal and conditional probability
functions and distribution functions in context of bivariate discrete random
variables. Similar functions, but in context of bivariate continuous random
variables, are discussed in this unit.
Bivariate continuous random variable is defined in Sec. 7.2. Joint and marginal
density functions are described in Sec. 7.3. Sec. 7.4 deals with the conditional
distribution and density functions. Independence of two continuous random
variables is dealt with in Sec. 7.5. Some practical problems on two-dimensional
continuous random variables are taken up in Sec. 7.6.
Objectives
A study of this unit would enable you to:
 define two-dimensional continuous random variable;
 specify the joint and marginal probability density functions of two
continuous random variables;
 obtain the conditional density and distribution functions for two-
dimensional continuous random variable;
 check the independence of two continuous random variables; and
 solve various practical problems on bivariate continuous random
variables.

41
Random Variables and 7.2 BIVARIATE CONTINUOUS RANDOM
Expectation
VARIABLES
Definition: If X and Y are continuous random variables defined on the sample
space S of a random experiment, then (X, Y) defined on the same sample space
S is called bivariate continuous random variable if (X, Y) assigns a point in
xy-plane defined on the sample space S. Notice that it (unlike discrete random
variable) assumes values in some non-countable set. Some examples of
bivariate continuous random variable are:
1. A gun is aimed at a certain point (say origin of the coordinate system).
Because of the random factors, suppose the actual hit point is any point
(X, Y) in a circle of radius unity about the origin.

1 X
-1
-1

Fig.7.1: Actual Hit Point when a Gun is Aimed at a Certain Point

Then (X, Y) assumes all the values in the circle  x, y  : x 2


 y 2  1 i.e.
(X, Y) assumes all values corresponding to each and every point in the
circular region as shown in Fig.7.1. Here, (X, Y) is bivariate continuous
random variable.

2. (X, Y) assuming all values in the rectangle  x, y  : a  x  b, c  y  d is a

bivariate continuous random variable.
Here, (X, Y) assumes all values corresponding to each and every point in
the rectangular region as shown in Fig.7.2.

y=d
d
x=a x=b
c
y=c
X
a b
Fig.7.2: (X, Y) Assuming All Values in the Rectangle  x,y  : a  x  b,c  y  d
3. In a statistical survey, let X denotes the daily number of hours a child
watches television and Y denotes the number of hours he/she spends on the
studies. Here, (X, Y) is a two-dimensional continuous random variable.

42
 Bivariate Continuous
7.3 JOINT AND MARGINAL DISTRIBUTION AND Random Variables
DENSITY FUNCTIONS
Two-Dimensional Continuous Distribution Function
The distribution function of a two-dimensional continuous random variable
(X, Y) is a real-valued function and is defined as
F  x, y   P  X  x, Y  y  for all real x and y.

Notice that the above function is in analogy with one-dimensional continuous

random variable case as studied in Unit 5 of the course.
Remark 1: F  x, y  can also be written as FX,Y  x, y  .

Joint Probability Density Function

Let (X, Y) be a continuous random variable assuming all values in some region
R of the xy-plane. Then, a function f  x, y  such that
x y

F  x, y     f  x, y  dydx
 

is defined to be a joint probability density function.

As in the one-dimensional case, a joint probability density function has the
following properties.
i) f  x, y   0
 
ii)   f  x, y  dydx  1
 

Remark 2:
As in the one-dimensional case, f  x, y  does not represent the probability of
anything. However, for positive x and y sufficiently small, f  x, y  xy is
approximately equal to
P  x  X  x  x, y  Y  y  y  .

In the one-dimensional case, you have studied that for positive x sufficiently
small f(x)x is approximately equal to P  x  X  x  x  . So, the two-
dimensional case is in analogy with the one-dimensional case.
Remark 3:
In analogy with the one-dimensional case [See Sec. 5.4 of Unit 5 of this course],
P  x  X  x  x, y  Y  y  y 
f  x, y  can be written as lim
x  0
y  0
xy

and is equal to
2
xy
 F  x, y   , i.e. second order partial derivative with respect to x and y.
43
 d
Random Variables and
Expectation
[See Sec. 5.5 of Unit 5 where f  x  
dx
 F  x  ]
2
Note:
xy
 F  x, y   means first differentiate F  x, y  partially w.r.t. y and
then the resulting function w.r.t. x. When we differentiate a function partially
w.r.t. one variable, then the other variable is treated as constant
For example, Let F  x, y   xy3  x 2 y

If we differentiate it partially w.r.t. y, we have


y
 F  x, y   = x(3y 2 )  x 2 .1 [ here, x is treated as constant.]

If we now differentiate this resulting expression w.r.t. x, we have

2
xy
 F  x, y   = 3 y 2  2 x [ here, y is treated as constant.]

Marginal Continuous Distribution Function

Let (X, Y) be a two-dimensional continuous random variable having f  x, y  as
its joint probability density function. Now, the marginal distribution function of
the continuous random variable X is defined as
F  x   P X  x 

 P  X  x, y    [ for X  x, Y can take any real value]

x
 
   f  x, y  dy  dx,
and the marginal distribution function of the continuous random variable Y is
defined as
F  y   P  Y  y

 P  Y  y, X    [ for Y  y, X can take any real value]

y
 
   f  x, y  dx  dy
Marginal Probability Density Functions
Let (X, Y) be a two-dimensional continuous random variable having
F  x, y  and f  x, y  as its distribution function and joint probability density
function, respectively. Let F  x  and F  y  be the marginal distribution
functions of X and Y, respectively. Then, the marginal probability density
function of X is given as

44
  Bivariate Continuous
f x   f  x, y  dy, Random Variables


d
or, it may also be obtained as
dx
 F  x  ,
and the marginal probability density function of Y is given as

f y   f  x, y  dx


or
d
=
dy
 F  y 

7.4 CONDITIONAL DISTRIBUTION AND

DENSITY FUNCTIONS
Conditional Probability Density Function
Let (X, Y) be a two-dimensional continuous random variable having the joint
probability density function f(x, y). The conditional probability density function
of Y given X = x is defined as
f  x, y 
f y  x  , where f(x) > 0 is the marginal density of X.
f x

Similarly, the conditional probability density function of X given Y = y is

defined to be
f  x, y 
f x  y  , where f(y) > 0 is the marginal density of Y.
f  y

As f  y  x  and f  x  y  , though conditional yet, are the probability density

functions, hence possess the properties of a probability density function.
Properties of f  y  x  are:

i) f  y  x  is clearly  0
 
f  x, y 
ii)  f  y  x  dy   f x dy
 


1  
   f  x, y  dy 
f  x    

  
1   f  x, y  dyis the m arg inal 

f x
f (x)   
 probability density function of X 

=1

45
Random Variables and Similarly, f  x  y  satisfies
Expectation
i) f  x  y   0 and

ii)  f  x  y  dx  1


Conditional Continuous Distribution Function

For a two-dimensional continuous random variable (X, Y), the conditional
distribution function of Y given X = x is defined as
F  y  x   P Y  y  X  x 
y

  f  y  x  dy , for all x such that f  x   0 ;



and the conditional distribution function of X given Y = y is defined as

F  x  y   P X  x  Y  y
x
  f  x  y  dx , for all y such that f(y) > 0.


7.5 STOCHASTIC INDEPENDENCE OF TWO

CONTINUOUS RANDOM VARIABLES
You have already studied in Unit 3 of this course that independence of events is
closely related to conditional probability, i.e. if events A and B are independent,
then P  A  B  P  A  , i.e. conditional probability of A is equal to the
unconditional probability of A. Likewise independence of random variables is
closely related to conditional distributions of random variables, i.e. two random
variables X and Y with joint probability function f  x, y  and marginal
probability functions f  x  and f  y  respectively are said to be stochastically
independent if and only if
i) f  y  x   f  y 

ii) f  x  y   f  x  .

Now, as defined in Sec. 7.4, we have

f  x, y 
f y  x 
f x

 f  x, y   f  x  f  y  x  [On cross-multiplying]

So, if X and Y are independent, then

f  x, y   f  x  f  y  [ f  y  x   f  y  ]

Remark 4: The random variables, if independent, are actually stochastically

46
independent but often the word “stochastically” is omitted.
Definition: Two random variables are said to be (stochastically) independent if Bivariate Continuous
and only if their joint probability density function is the product of their Random Variables
marginal density functions.
Let us now take up some problems on the topics covered so far in this unit.

7.6 PROBLEMS ON TWO-DIMENSIONAL

CONTINUOUS RANDOM VARIABLES

Example 1: Let X and Y be two random variables. Then for

k  2x  y  , 0  x  1, 0  y  2
f  x, y   
 0, elsewhere

to be a joint density function, what must be the value of k ?

Solution: As f  x, y  is the joint probability density function,
 
   f  x, y  dy dx  1
 

1 2
   f  x, y  dy dx  1
0 0
[0  x  1, 0  y  2 ]

1 2
   k  2x  y  dy dx  1
0 0

1 2
 
 k     2x  y  dy  dx  1
0 0 
1 2
 y2 
 k   2xy   dx  1
0 
2 0

[Firstly the integral has been done w.r.t. y treating x as constant.]

1  2

 k   2x  2  
 2 
 0  dx  1
0 
2 

1
 k   4x  2  dx  1
0

1
 4x 2 
 k  2x   1
 2 0
4  1
 k   2  0   1  4k  1  k =
2  4

47
Random Variables and
Example 2: Let the joint density function of a two-dimensional random
Expectation variable (X, Y) be:
x  y for 0  x  1 and 0  y  1
f  x, y   
 0, otherwise

Find the conditional density function of Y given X.

f  x, y 
Solution: The conditional density function of Y given X is f  y  x   ,
f x

where f  x, y  is the joint density function, which is given; and f  x  is the

marginal density function which, by definition, is given by

f x   f  x, y  dy


1
  f  x, y  dy [ 0  y < 1]
0

1
=   x  y  dy
0

1
 y2 
  xy   .
 2 0
2
 1  1
  x 1   0   x  , 0  x  1.
 2  2

 the conditional density function of Y given X is

f  x, y  xy
f y  x   , for 0  x < 1 and 0  y < 1.
f x x
1
2
Example 3: Two-dimensional random variable (X, Y) have the joint density
8xy, 0  x  y  1
f  x, y   
 0 , otherwise
1 1
i) Find P[X <  Y < ].
2 4
ii) Find the marginal and conditional distributions.
iii)Are X and Y independent?
Solution:
1 1 1 1 1 1

 1 1 2 4 2 4 2
 y2  4
i) P  X   Y   f  x, y dy dx  8xy dy dx  0  2  dx
8x
 2 4  0 0 0 0
0

48
 1 1 1 Bivariate Continuous
2  1  2
x 1  x2  2 Random Variables
  8x  dx  0 4 dx   
0 16  2   4  2 0

1 1  1
  .
4  8  32

ii) Marginal density function of X is

1
f  x    f  x, y  dy [0  x  y  1]
x

1 1
 y2 
  8xy dy  8x  
x  2 x

 1 x2 

 8x     4x 1  x 2 for 0  x  1 
2 2 
Marginal density function of Y is
y

f  y    f  x, y  dx [0  x  y ]
0

  8xy dx
0

y
 x2  8y 3
 8y     4y 3 for 0  y  1
 2 0 2

Conditional density function of X given Y(0 < Y < 1) is

f  x, y 
f  x  y 
f  y

8xy 2x
  , 0x y
4y 3 y 2
Conditional density function of Y given X(0 < X < 1) is
f  x, y 
f y  x 
f x

8xy 2y
  , x < y <1

4x 1  x 2   1 x2 
iii) f  x, y   8xy,

 
But f  x  f  y   4x 1  x 2 4y3

 16x 1  x  y
2 3

49
Random Variables and  f  x, y   f  x  f  y 
Expectation
Hence, X and Y are not independent random variables.
Now, you can try some exercises.
E1) Let X and Y be two random variables. Then for
kxy for 0  x  4 and 1  y  5
f  x, y   
 0, otherwise
to be a joint density function, what must be the value of k?
E2) If the joint p.d.f. of a two-dimensional random variable (X, Y) is given by
2 for 0  x  1 and 0  y  x
f  x, y   
0, otherwise,
Then,
i) Find the marginal density functions of X and Y.
ii) Find the conditional density functions.
iii) Check for independence of X and Y.
E3) If (X, Y) be two-dimensional random variable having joint density
function.

1
  6  x  y ; 0  x  2 , 2  y  4
f  x, y    8
 0, elsewhere

Find (i) P  X  1, Y  3 (ii) P  X  1  Y  3

Now before ending this unit, let’s summarize what we have covered in it.

7.7 SUMMARY
In this unit, we have covered the following main points:
1) If X and Y are continuous random variables defined on the sample space S of
a random experiment, then (X, Y) defined on the same sample space S is
called bivariate continuous random variable if (X, Y) assigns a point in
xy -plane defined on the sample space S.
2) The distribution function of a two-dimensional continuous random variable
(X, Y) is a real-valued function and is defined as
F  x, y   P  X  x, Y  y  for all real x and y.

3) A function f  x, y  is called joint probability density function if it is such

that
x y

F  x, y     f  x, y  dydx
 

50
 and satisfies Bivariate Continuous
Random Variables
i) f  x, y   0
 
ii)   f  x, y  dydx  1.
 
4) The marginal distribution function of the continuous random variable X is
defined as
x
 
F  x   P X  x      f  x, y  dy dx,
and that of continuous random variable Y is defined as
y
 
F  y   P  Y  y    f  x, y  dx dy .
5) The marginal probability density function of X is given as

d
f x   f  x, y  dy  dx  F(x)  ,


and that of Y is given as


d
f y   f  x, y  dx = dx
 F  y  .


6) The conditional probability density function of Y given X = x is defined

as
f  x, y 
f y  x  ,
f x

and that of X given Y = y is defined as

f  x, y 
f x  y  .
f  y

7) The conditional distribution function of Y given X = x is defined as

y

F y  x   f  y  x  dy , for all x such that f  x   0 ;



and that of X given Y = y is defined as

x
F  x  y   f  x  y  dx , for all y such that f(y) > 0.


8) Two random variables are said to be (stochastically) independent if and

only if their joint probability density function is the product of their
marginal density functions.

51
Random Variables and 7.8 SOLUTIONS/ANSWERS
Expectation
E1) As f  x, y  is the joint probability density function,
 
   f  x, y dy dx  1
 

4 5 4 5
 
 0 1 kxy dy dx  1  k 0  1 xy dy  dx  1
4 5 4
 y2 
 k   x  dx  1  k  12x dx  1
0 
2 1 0

4
 x2 
 12k    1  96 k = 1
 2 0
1
k=
96
E2) i) Marginal density function of Y is given by
 1

f y   f  x, y  dx   2dx
 y

[As x is involved in both the given ranges, i.e. 0 < x < 1 and 0 < y < x;
therefore, here we will combine both these intervals and hence have
0 < y < x < 1.  x takes the values from y to 1]
1
  2x y  2  2y

= 2 – 2y
= 2(1– y), 0 < y < 1
Marginal density function of X is given by

f x   f  x, y  dy


x
  2dy [ 0 < y < x < 1]
0

x
 2  y 0
 2x, 0  x  1.
ii) Conditional density function of Y given X(0 < X < 1) is
f  x, y  2 1
f y  x    ; 0  y x
f x 2x x

Conditional density function of X and given Y(0 < Y < 1) is

f  x, y  2 1
f  x  y    , y x 1
f  y 2 1  y  1  y
52
 iii) f  x, y   2, Bivariate Continuous
Random Variables

f  x  f  y   2  2x 1  y 

As f  x, y   f  x  f  y  ,

 X and Y are not independent.

1 3
E3) (i) P  X  1, Y  3    f  x, y  dy dx
 

1 3
1
   6  x  y dy dx
0 2
8
3
1
1  y2 
    6y  xy    dx
8
0  
2  2
1
1  9 
   6  3   x  3    12  2x  2dx
8 0  2 
1
1  9 
   18  3x    10  2x  dx
8 0  2 
1 1
1 7  1 7 x2  1 7 1  3
    x  dx   x       
802  8 2 2 0 8  2 2  8

P  X  1, Y  3
ii) P  X  1  Y  3 
P  Y  3
2 3
1
where P  Y  3      6  x  y  dy dx
0 2
8
2 3
1  y2 
  6y  xy   dx .
80 2 2
2
1  9 
   18  3x    12  2x  2 dx
8 0  2 
2
1  9 
   18  3x    10  2x   dx
8 0  2 
2
1 7 
   x  dx
8 0  2 
2
1 7 x2 
  x 
8 2 2 0

53
Random Variables and 1 4 
 7   0
Expectation 8 2 
5

8
3 / 8  value of numerator is 
 P  X  1 Y  3 
5 / 8 already calculated in part(i) 


3

5

54