Module 21 Independent Random Variables

Module 21
INDEPENDENT RANDOM VARIABLES
() Module 21 INDEPENDENT RANDOM VARIABLES 1 / 11

X = (X1 , · · · , Xp ) : a p-dimensional r.v. with joint d.f. FX (·);
FXi (·) : marginal d.f. of Xi , i = 1, · · · , p, i.e.,
FXi (x) = x lim F (x , · · · , xi−1 , x, xi+1 , · · · , xp ), i = 1, · · · , p.

→∞ X 1
j
j6=i

Definition 1:
(a) Random variables X1 , . . . , Xp are said to be independent if, for any

subcollection {Xλ1 , . . . , Xλq } ⊆ {X1 , . . . , Xp } (2 ≤ q ≤
p, {λ1 , . . . , λq } ⊆ {1, . . . , p}), we have
q
Y
FXλ1 ,...,Xλq (x1 , . . . , xq ) = FXλi (xi ), ∀ x = (x1 , . . . , xq ) ∈ Rq .
i=1
(b) Let {Xλ : λ ∈ Λ} be a family of random variables. The r.v.s in the

family {Xλ : λ ∈ Λ} are said to be independent if those in any finite
subcollection of {Xλ : λ ∈ Λ} are independent.

Remark 1:
(a) If Λ1 ⊆ Λ2 and the r.v.s {Xλ : λ ∈ Λ2 } are independent then r.v.s

{Xλ : λ ∈ Λ1 } are also independent.
(b) The above definition of independent r.v.s can be extended to
independence of random vectors (of possibly different dimensions) in
an obvious manner. For example, let X = (X1 , . . . , Xp ) be a
p-dimensional r.v. with joint d.f. F (·), Y = (Y1 , . . . , Yq ) is a
q-dimensional r.v. with joint d.f. G (·), and let
Z = (X1 , . . . , Xp , Y1 , . . . , Yq ) = (X , Y ) (a (p + q)-dimensional r.v.)
have joint d.f. H(·). Then X and Y are said to be independent iff
H(x, y ) = F (x)G (y ), ∀ z = (x, y ) ∈ Rp+q .

Result 1 :
R.V.s X1 , . . . Xp are independent iff
p
Y
FX1 ,...,Xp (x1 , . . . , xp ) = FXi (xi ), ∀ x = (x1 , . . . , xp ) ∈ Rp . (1)
i=1
Proof: Suppose that X1 , . . . , Xp are independent. Then, by definition, (1)

holds.
Conversely suppose that (1) holds. Then, for x = (x1 , . . . , xp−1 ) ∈ Rp−1 ,
FX1 ,...,Xp−1 (x1 , . . . , xp−1 ) = lim FX1 ,...,Xp−1 ,Xp (x1 , . . . , xp−1 , xp )
xp →∞
p
Y
= lim FXj (xj )
xp →∞
j=1
p−1
Y
= lim FXp (xp ) FXj (xj )
xp →∞
j=1

p−1
Y
= FXj (xj ).
j=1
In general, for 2 ≤ r ≤ p, {Xλ1 , . . . , Xλr } ⊆ {X1 , . . . , Xp } and

x = (x1 , . . . , xr ) ∈ Rr
r
Y
FXλ1 ,...,Xλr (x1 , . . . , xr ) = FXλi (xi ).
i=1
Remark 2: The above results remain valid if random variables X1 , . . . , Xp

are replaced by random vectors X 1 , . . . , X p of (possibly) different
dimensions.

Example 1: Let (X , Y ) be a r.v. with joint d.f.


 0, if x < 0 or y < 0

2 0 ≤ y < 1 and x ≥ y
y , if



FX ,Y (x, y ) = x(2y − x), if 0≤x <y <1 .

x(2 − x), if 0 ≤ x < 1, y ≥ 1





1, if x ≥ 1, y ≥ 1

The marginal d.f.s of X and Y are


0,
 if x < 0
FX (x) = lim FX ,Y (x, y ) = x(2 − x), if 0 ≤ x < 1 ,
y →∞ 
1, if x ≥ 1

and 
0,
 if y < 0
FY (y ) = lim FX ,Y (x, y ) = y 2 , if 0 ≤ y < 1 ,
x→∞ 
1, if y ≥ 1.

respectively.
Clearly,
FX ,Y (x, y ) 6= FX (x)FY (y ), ∀ (x, y ) ∈ R2
⇒ X and Y are not independent.
Example 2: Let the r.v. (X , Y ) have the joint d.f.



 0, if y < 0




 x
)e−y )
FX ,Y (x, y ) = e (1−(1+y
2 , if x < 0, y ≥ 0 .





 (2−e−x )(1−(1+y )e−y )

2 , if x ≥ 0, y ≥ 0
The marginal d.f.s of X and Y are
(
ex
FX (x) = lim FX ,Y (x, y ) = 2, if x < 0
2−e−x
y →∞
2 , if x ≥ 0
and (
0, if y < 0
FY (y ) = lim FX ,Y (x, y ) = −y
,
x→∞ 1 − (1 + y )e , if y ≥ 0
respectively. Clearly X and Y are independent, as
FX ,Y (x, y ) = FX (x)FY (y ), ∀ (x, y ) ∈ R2 .

Take Home Problems
Let F and G be d.f.s on R. Define,
H1 (x, y ) = min{F (x), G (y )}, (x, y ) ∈ R2

and
H2 (x, y ) = max{F (x) + G (y ) − 1, 0}, (x, y ) ∈ R2 .
(a) Show that H1 and H2 are d.f.s on R2 .
(b) Find marginal d.f.s of H1 . Check independence.
(c) Find marginal d.f.s of H2 . Check independence.

Abstract of Next Module
We will introduce discrete random vectors and study their probability

distributions.

Thank you for your patience

Module 21 Independent Random Variables

Uploaded by

Copyright:

Available Formats

Module 21 Independent Random Variables

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Module 21 Independent Random Variables

Uploaded by

Copyright:

Available Formats

Module 21

INDEPENDENT RANDOM VARIABLES

() Module 21 INDEPENDENT RANDOM VARIABLES 1 / 11

FXi (·) : marginal d.f. of Xi , i = 1, · · · , p, i.e.,

FXi (x) = x lim F (x , · · · , xi−1 , x, xi+1 , · · · , xp ), i = 1, · · · , p.

() Module 21 INDEPENDENT RANDOM VARIABLES 1 / 11

(a) Random variables X1 , . . . , Xp are said to be independent if, for any

(b) Let {Xλ : λ ∈ Λ} be a family of random variables. The r.v.s in the

() Module 21 INDEPENDENT RANDOM VARIABLES 2 / 11

(a) If Λ1 ⊆ Λ2 and the r.v.s {Xλ : λ ∈ Λ2 } are independent then r.v.s

H(x, y ) = F (x)G (y ), ∀ z = (x, y ) ∈ Rp+q .

() Module 21 INDEPENDENT RANDOM VARIABLES 3 / 11

Proof: Suppose that X1 , . . . , Xp are independent. Then, by definition, (1)

() Module 21 INDEPENDENT RANDOM VARIABLES 4 / 11

In general, for 2 ≤ r ≤ p, {Xλ1 , . . . , Xλr } ⊆ {X1 , . . . , Xp } and

Remark 2: The above results remain valid if random variables X1 , . . . , Xp

() Module 21 INDEPENDENT RANDOM VARIABLES 5 / 11

The marginal d.f.s of X and Y are

⇒ X and Y are not independent.

Example 2: Let the r.v. (X , Y ) have the joint d.f.

FX ,Y (x, y ) = FX (x)FY (y ), ∀ (x, y ) ∈ R2 .

() Module 21 INDEPENDENT RANDOM VARIABLES 8 / 11

Let F and G be d.f.s on R. Define,

H1 (x, y ) = min{F (x), G (y )}, (x, y ) ∈ R2

() Module 21 INDEPENDENT RANDOM VARIABLES 9 / 11

We will introduce discrete random vectors and study their probability

() Module 21 INDEPENDENT RANDOM VARIABLES 10 / 11

() Module 21 INDEPENDENT RANDOM VARIABLES 11 / 11

You might also like