0% found this document useful (0 votes)
70 views4 pages

3 Bivariate Transformations

This document discusses bivariate transformations of random vectors. It defines how transforming two random variables (X,Y) into new random variables (U,V) via functions g1(x,y) and g2(x,y) affects their joint and marginal distributions. For discrete (X,Y), the distribution of (U,V) is determined by the probabilities of (X,Y) falling within regions mapped by the transformation. For continuous (X,Y), the joint distribution of (U,V) involves the Jacobian of the transformation. The document provides examples of how specific bivariate transformations distribute sums, ratios, and other functions of random variables.

Uploaded by

Jiah Mohd Daud
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
70 views4 pages

3 Bivariate Transformations

This document discusses bivariate transformations of random vectors. It defines how transforming two random variables (X,Y) into new random variables (U,V) via functions g1(x,y) and g2(x,y) affects their joint and marginal distributions. For discrete (X,Y), the distribution of (U,V) is determined by the probabilities of (X,Y) falling within regions mapped by the transformation. For continuous (X,Y), the joint distribution of (U,V) involves the Jacobian of the transformation. The document provides examples of how specific bivariate transformations distribute sums, ratios, and other functions of random variables.

Uploaded by

Jiah Mohd Daud
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

3 Bivariate Transformations

Let (X, Y ) be a bivariate random vector with a known probability distribution. Let U = g
1
(X, Y )
and V = g
2
(X, Y ), where g
1
(x, y) and g
2
(x, y) are some specied functions. If B is any subset of
R
2
, then (U, V ) B if and only if (X, Y ) A, where A = {(x, y) : (g
1
(x, y), g
2
(x, y)) B}. Thus
P((U, V ) B) = P((X, Y ) A), and the probability of (U, V ) is completely determined by the
probability distribution of (X, Y ).
If (X, Y ) is a discrete bivariate random vector, then
f
U,V
(u, v) = P(U = u, V = v) = P((X, Y ) A
u,v
) =

(x,y)Auv
f
X,Y
(x, y),
where A
u,v
= {(x, y) : g
1
(x, y) = u, g
2
(x, y) = v}.
Example 3.1 (Distribution of the sum of Poisson variables) Let X and Y be independent Poisson
random variables with parameters and , respectively. Thus, the joint pmf of (X, Y ) is
f
X,Y
(x, y) =

x
e

x!

y
e

y!
, x = 0, 1, 2, . . . , y = 0, 1, 2, . . .
Now dene U = X +Y and V = Y , thus,
f
U,V
(u, v) = f
X,V
(u v, v) =

uv
e

(u v)!

v
e

v!
, v = 0, 1, 2, . . . , u = v, v + 1, . . .
The marginal of U is
f
U
(u) =
u

v=0

uv
e

(u v)!

v
e

v!
= e
(+)
u

v=0

uv
(u v)!

v
v!
=
e
(+)
u!
u

v=0

u
v

uv
=
e
(+)
u!
( +)
u
, u = 0, 1, 2, . . .
This is the pmf of a Poisson random variable with parameter +.
Theorem 3.1 If X Poisson() and Y Poisson() and X and Y are independent, then
X +Y Poisson( +).
If (X, Y ) is a continuous random vector with joint pdf f
X,Y
(x, y), then the joint pdf of (U, V )
can be expressed in terms of F
X,Y
(x, y) in a similar way. As before, let A = {(x, y) : f
X,Y
(x, y) > 0}
and B = {(u, v) : u = g
1
(x, y) and v = g
2
(x, y) for some (x, y) A}. For the simplest version of
this result, we assume the transformation u = g
1
(x, y) and v = g
2
(x, y) denes a one-to-one
transformation of A to B. For such a one-to-one, onto transformation, we can solve the equations
10
u = g
1
(x, y) and v = g
2
(x, y) for x and y in terms of u and v. We will denote this inverse
transformation by x = h
1
(u, v) and y = h
2
(u, v). The role played by a derivative in the univariate
case is now played by a quantity called the Jacobian of the transformation. It is dened by
J =

x
u
x
v
y
u
y
v

,
where
x
u
=
h
1
(u,v)
u
,
x
v
=
h
1
(u,v)
v
,
y
u
=
h
2
(u,v)
u
, and
y
v
=
h
2
(u,v)
v
.
We assume that J is not identically 0 on B. Then the joint pdf of (U, V ) is 0 outside the set B
and on the set B is given by
f
U,V
(u, v) = f
X,Y
(h
1
(u, v), h
2
(u, v))|J|,
where |J| is the absolute value of J.
Example 3.2 (Sum and dierence of normal variables) Let X and Y be independent, standard
normal variables. Consider the transformation U = X + Y and V = X Y . The joint pdf of X
and Y is, of course,
f
X,Y
(x, y) = (2)
1
exp(x
2
/2) exp(y
2
/2), < x < , < y < .
so the set A = R
2
. Solving the following equations
u = x +y and v = x y
for x and y, we have
x = h
1
(x, y) =
u +v
2
, and y = h
2
(x, y) =
u v
2
.
Since the solution is unique, we can see that the transformation is one-to-one, onto transformation
from A to B = R
2
.
J =

x
u
x
v
y
u
y
v

1
2
1
2
1
2

1
2

=
1
2
.
So the joint pdf of (U, V ) is
f
U,V
(u, v) = f
X,Y
(h
1
(u, v), h
2
(u, v))|J| =
1
2
e
((u+v)/2)
2
/2
e
((uv)/2)
2
/2
1
2
for < u < and < v < . After some simplication and rearrangement we obtain
f
U,V
(u, v) = (
1

2p

2
e
u
2
/4
)(
1

2p

2
e
v
2
/4
).
The joint pdf has factored into a function of u and a function of v. That implies U and V are
independent.
11
Theorem 3.2 Let X and Y be independent random variables. Let g(x) be a function only of x and
h(y) be a function only of y. Then the random variables U = g(X) and V = h(Y ) are independent.
Proof: We will prove the theorem assuming U and V are continuous random variables. For any
u mR and v R, dene
A
u
= {x : g(x) u} and B
u
= {y : h(y) v}.
Then the joint cdf of (U, V ) is
F
U,V
(u, v) = P(U u, V v)
= P(X A
u
, Y B
v
)
P(X A
u
)P(Y B
v
).
The joint pdf of (U, V ) is
f
U,V
(u, v) =

2
uv
F
U,V
(u, v) = (
d
du
P(X A
u
))(
d
dv
P(Y B
v
)),
where the rst factor is a function only of u and the second factor is a function only of v. Hence,
U and V are independent.

In many situations, the transformation of interest is not one-to-one. Just as Theorem 2.1.8
(textbook) generalized the univariate method to many-to-one functions, the same can be done
here. As before, A = {(x, y) : f
X,Y
(x, y) > 0}. Suppose A
0
, A
1
, . . . , A
k
form a partition of A
with these properties. The set A
0
, which may be empty, satises P((X, Y ) A
0
) = 0. The
transformation U = g
1
(X, Y ) and V = g
2
(X, Y ) is a one-to-one transformation from A
i
onto B for
each i = 1, 2, . . . , k. Then for each i, the inverse function from B to A
i
can be found. Denote the
ith inverse by x = h
1i
(u, v) and y = h
2i
(u, v). Let J
i
denote the Jacobian computed from the ith
inverse. Then assuming that these Jacobians do not vanish identically on B, we have
f
U,V
(u, v) =
k

i=1
f
X,Y
(h
1i
(u, v), h
2i
(u, v))|J
i
|.
Example 3.3 (Distribution of the ratio of normal variables) Let X and Y be independent N(0, 1)
random variable. Consider the transformation U = X/Y and V = |Y |. (U and V can be dened
to be any value, say (1,1), if Y = 0 since P(Y = 0) = 0.) This transformation is not one-to-one,
since the points (x, y) and (x, y) are both mapped into the same (u, v) point. Let
A
1
= {(x, y) : y > 0}, A
2
= {(x, y) : y < 0}, A
0
= {(x, y) : y = 0}.
12
A
0
, A
1
and A
2
form a partition of A = R
2
and P(A
0
) = 0. The inverse transformations from B
to A
1
and B to A
2
are given by
x = h
11
(u, v) = uv, y = h
21
(u, v) = v,
and
x = h
12
(u, v) = uv, y = h
22
(u, v) = v.
The Jacobians from the two inverses are J
1
= J
2
= v. Using
f
X,Y
(x, y) =
1
2
e
x
2
/2
e
y
2
/2
,
we have
f
U,V
(u, v) =
1
2
e
(uv)
2
/2
e
v
2
/2
|v| +
1
2
e
(uv)
2
/2
e
(v)
2
/2
|v|
=
v

e
(u
2
+1)v
2
/2
, < u < , 0 < v < .
From this the marginal pdf of U can be computed to be
f
U
(u) =


0
v

e
(u
2
+1)v
2
/2
dv
=
1
2


0
e
(u
2
+1)z/2
dz (z = v
2
)
=
1
(u
2
+ 1)
So we see that the ratio of two independent standard normal random variable is a Cauchy random
variable.
13

You might also like