Chapter 3.

Vectors of Random Variables

Nguyễn Văn Hạnh

Department of Applied Mathematics
School of Applied Mathematics and Informatics

First semester, 2023-2024

NV HANH Probability and statistics First semester, 2023-2024 1 / 39

Contents

1 Joint probability distrubutions

2 Marginal probability distribution

3 Conditional probability distribution

4 Independence

5 Examples

6 Covariance matrix and correlation

7 Examples

NV HANH Probability and statistics First semester, 2023-2024 2 / 39

 Joint probability distrubutions

Introduction

In practice, it is often necessary to consider some different variables at

the same time that are related and lead to the concept of a vector of
random variables.

NV HANH Probability and statistics First semester, 2023-2024 3 / 39

 Joint probability distrubutions

Introduction

In practice, it is often necessary to consider some different variables at

the same time that are related and lead to the concept of a vector of
random variables.
Studying each variable separately may give incomplete information.

NV HANH Probability and statistics First semester, 2023-2024 3 / 39

 Joint probability distrubutions

Introduction

In practice, it is often necessary to consider some different variables at

the same time that are related and lead to the concept of a vector of
random variables.
Studying each variable separately may give incomplete information.
For example: when we observed the characteristics of a machine, we
are interested in some different variables at the same time, such as
the weight, size, quality, material, . . .

NV HANH Probability and statistics First semester, 2023-2024 3 / 39

 Joint probability distrubutions

Introduction

For simplicity, we study two-dimensional random variables (X , Y ),

where X and Y are one-dimensional variables.

NV HANH Probability and statistics First semester, 2023-2024 4 / 39

 Joint probability distrubutions

Introduction

For simplicity, we study two-dimensional random variables (X , Y ),

where X and Y are one-dimensional variables.
Most results can be extended easily to n-dimensional variables.

NV HANH Probability and statistics First semester, 2023-2024 4 / 39

 Joint probability distrubutions

Introduction

For simplicity, we study two-dimensional random variables (X , Y ),

where X and Y are one-dimensional variables.
Most results can be extended easily to n-dimensional variables.
This chapter analyzes experiments that produce two random
variables, X and Y.

NV HANH Probability and statistics First semester, 2023-2024 4 / 39

 Joint probability distrubutions

Introduction

For simplicity, we study two-dimensional random variables (X , Y ),

where X and Y are one-dimensional variables.
Most results can be extended easily to n-dimensional variables.
This chapter analyzes experiments that produce two random
variables, X and Y.
If X and Y are discrete, we have a discrete two-dimensional random
variable; if they are continuous, we have a continuous
two-dimensional variable.

NV HANH Probability and statistics First semester, 2023-2024 4 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.1:
Let X and Y be two discrete random variables, where the range of X is
SX = {x1 , x2 , .., xn } and the range of Y is SY = {y1 , y2 , .., ym }.

NV HANH Probability and statistics First semester, 2023-2024 5 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.1:
Let X and Y be two discrete random variables, where the range of X is
SX = {x1 , x2 , .., xn } and the range of Y is SY = {y1 , y2 , .., ym }.
The range of two-dimensional random variable (X , Y ) is
SX ,Y = {(xi , yj ) : i = 1, ..., n; j = 1, ..., m}.

NV HANH Probability and statistics First semester, 2023-2024 5 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.1:
Let X and Y be two discrete random variables, where the range of X is
SX = {x1 , x2 , .., xn } and the range of Y is SY = {y1 , y2 , .., ym }.
The range of two-dimensional random variable (X , Y ) is
SX ,Y = {(xi , yj ) : i = 1, ..., n; j = 1, ..., m}.
The joint probability distrubution of (X , Y ) is defined by the joint
probability mass function (pmf):
(
pij if x = xi , y = yj ,
f (x, y ) = P(X = x, Y = y ) =
0 otherwise
Pn Pm
such that pij ≥ 0∀i, j and i=1 j=1 pij = 1.

NV HANH Probability and statistics First semester, 2023-2024 5 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.1:
The joint probability distrubution of two-dimensional discrete random
variable (X , Y ) is also given by the following table:
y1 y2 ... yj ... ym
x1 p11 p12 ... p1j ... p1m
x2 p21 p22 ... p2j ... p2m
. . . ... . ... .
. . . ... . ... .
. . . ... . ... .
xi pi1 pi2 ... pij ... pim
. . . ... . ... .
. . . ... . ... .
. . . ... . ... .
xn pn1 pn2 ... pnj ... pnm

NV HANH Probability and statistics First semester, 2023-2024 6 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.2:
Let X and Y be two continuous random variables. The joint probability
distrubution of (X , Y ) is defined by the joint probability density function
(pdf) f (x, y ) that satisfies the following requirements:

NV HANH Probability and statistics First semester, 2023-2024 7 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.2:
Let X and Y be two continuous random variables. The joint probability
distrubution of (X , Y ) is defined by the joint probability density function
(pdf) f (x, y ) that satisfies the following requirements:
For all (x, y ) ∈ R 2 : f (x, y ) ≥ 0.

NV HANH Probability and statistics First semester, 2023-2024 7 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.2:
Let X and Y be two continuous random variables. The joint probability
distrubution of (X , Y ) is defined by the joint probability density function
(pdf) f (x, y ) that satisfies the following requirements:
For all (x, y ) ∈ R 2 : f (x, y ) ≥ 0.
R +∞ R +∞
−∞ −∞ f (x, y )dxdy = 1.

NV HANH Probability and statistics First semester, 2023-2024 7 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.2:
Let X and Y be two continuous random variables. The joint probability
distrubution of (X , Y ) is defined by the joint probability density function
(pdf) f (x, y ) that satisfies the following requirements:
For all (x, y ) ∈ R 2 : f (x, y ) ≥ 0.
R +∞ R +∞
−∞ −∞ f (x, y )dxdy = 1.
For D ⊂ R 2 : P[(X , Y ) ∈ D] =
RR
D f (x, y )dxdy .

NV HANH Probability and statistics First semester, 2023-2024 7 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.3:
Let (X , Y ) be a two-dimensional random variable. The joint cumulative
distribution function (cdf) of (X , Y ) is defined by:

F (x, y ) = P(X < x, Y < y ), ∀(x, y ) ∈ R 2 .

NV HANH Probability and statistics First semester, 2023-2024 8 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.3:
Let (X , Y ) be a two-dimensional random variable. The joint cumulative
distribution function (cdf) of (X , Y ) is defined by:

F (x, y ) = P(X < x, Y < y ), ∀(x, y ) ∈ R 2 .

P P
If (X , Y ) is discrete then F (x, y ) = xi <x yj <y P(X = xi , Y = yj ).

NV HANH Probability and statistics First semester, 2023-2024 8 / 39

 Joint probability distrubutions

Joint probability distrubutions

Definitions 4.3:
Let (X , Y ) be a two-dimensional random variable. The joint cumulative
distribution function (cdf) of (X , Y ) is defined by:

F (x, y ) = P(X < x, Y < y ), ∀(x, y ) ∈ R 2 .

P P
If (X , Y ) is discrete then F (x, y ) = xi <x yj <y P(X = xi , Y = yj ).
Rx Ry
If (X , Y ) is continuous then F (x, y ) = −∞ −∞ f (u, v )dudv .

NV HANH Probability and statistics First semester, 2023-2024 8 / 39

 Marginal probability distribution

Marginal probability distribution

Definitions 4.4:
Let (X , Y ) be a two-dimensional discrete random variable with the joint
pmf (
pij if (x, y ) = (xi , yj ),
f (x, y ) = P(X = x, Y = y ) =
0 otherwise

NV HANH Probability and statistics First semester, 2023-2024 9 / 39

 Marginal probability distribution

Marginal probability distribution

Definitions 4.4:
Let (X , Y ) be a two-dimensional discrete random variable with the joint
pmf (
pij if (x, y ) = (xi , yj ),
f (x, y ) = P(X = x, Y = y ) =
0 otherwise

The marginal probability distribution of X is the defined by:

X x1 x2 ... xi ... xn
P p1· p2· ... pi· ... pn·
Pm
where pi· = j=1 pij , i = 1, .., n.

NV HANH Probability and statistics First semester, 2023-2024 9 / 39

 Marginal probability distribution

Marginal probability distribution

Definitions 4.4:
Let (X , Y ) be a two-dimensional discrete random variable with the joint
pmf (
pij if (x, y ) = (xi , yj ),
f (x, y ) = P(X = x, Y = y ) =
0 otherwise

The marginal probability distribution of X is the defined by:

X x1 x2 ... xi ... xn
P p1· p2· ... pi· ... pn·
Pm
where pi· = j=1 pij , i = 1, .., n.
The marginal probability distribution of Y is the defined by:
Y y1 y2 ... yj ... ym
P p·1 p·2 ... p·j ... p·m
Pn
where p·j = i=1 pij , j = 1, .., m.
NV HANH Probability and statistics First semester, 2023-2024 9 / 39
 Marginal probability distribution

Marginal probability distrubutions

Definitions 4.4:
The marginal probability distrubution of X and of Y :
y1 y2 ... yj ... ym P(xi )
x1 p11 p12 ... p1j ... p1m p1·
x2 p21 p22 ... p2j ... p2m p2·
. . . ... . ... . .
. . . ... . ... . .
. . . ... . ... . .
xi pi1 pi2 ... pij ... pim pi·
. . . ... . ... . .
. . . ... . ... . .
. . . ... . ... . .
xn pn1 pn2 ... pnj ... pnm pn·
P(yj ) p·1 p·2 ... p·j ... p·m 1

NV HANH Probability and statistics First semester, 2023-2024 10 / 39

 Marginal probability distribution

Marginal probability distribution

Definitions 4.5:
Let (X , Y ) be a two-dimensional continuous random variable with the
joint pdf f (x, y ).
The marginal probability distribution of X is the defined by the
marginal pdf: Z +∞
fX (x) = f (x, y )dy .
−∞

NV HANH Probability and statistics First semester, 2023-2024 11 / 39

 Marginal probability distribution

Marginal probability distribution

Definitions 4.5:
Let (X , Y ) be a two-dimensional continuous random variable with the
joint pdf f (x, y ).
The marginal probability distribution of X is the defined by the
marginal pdf: Z +∞
fX (x) = f (x, y )dy .
−∞

The marginal probability distribution of Y is the defined by the

marginal pdf: Z +∞
fY (y ) = f (x, y )dx.
−∞

NV HANH Probability and statistics First semester, 2023-2024 11 / 39

 Conditional probability distribution

Conditional probability distribution

Definitions 4.6:
Let (X , Y ) be a two-dimensional discrete random variable with the joint
pmf (
pij if (x, y ) = (xi , yj ),
f (x, y ) = P(X = x, Y = y ) =
0 otherwise

NV HANH Probability and statistics First semester, 2023-2024 12 / 39

 Conditional probability distribution

Conditional probability distribution

Definitions 4.6:
Let (X , Y ) be a two-dimensional discrete random variable with the joint
pmf (
pij if (x, y ) = (xi , yj ),
f (x, y ) = P(X = x, Y = y ) =
0 otherwise

The conditional probability distribution of X given that Y = yj is

defined by:
X |Y = yj x1 x2 ... xi ... xn
P p1j /p·j p2j /p·j ... pij /p·j ... pnj /p·j
since
P(X = xi , Y = yj ) pij
P(X = xi |Y = yj ) = = , i = 1, .., n
P(Y = yj ) p·j
.
NV HANH Probability and statistics First semester, 2023-2024 12 / 39
 Conditional probability distribution

Conditional probability distribution

Definitions 4.7:
Let (X , Y ) be a two-dimensional discrete random variable with the joint
pmf (
pij if (x, y ) = (xi , yj ),
f (x, y ) = P(X = x, Y = y ) =
0 otherwise

The conditional probability distribution of Y given that X = xi is

defined by:
Y |X = xi y1 y2 ... yj ... ym
P pi1 /pi· pi2 /pi· ... pij /pi· ... pim /pi·
since
P(X = xi , Y = yj ) pij
P(Y = yj |X = xi ) = = , j = 1, .., m
P(X = xi ) pi·
.
NV HANH Probability and statistics First semester, 2023-2024 13 / 39
 Conditional probability distribution

Conditional probability distribution

Definitions 4.7:
Let (X , Y ) be a two-dimensional continuous random variable with the
joint pdf f (x, y ).
The conditional probability distribution of X givent that Y = y is
defined by conditional pdf:

f (x, y )
fX |Y =y (x) = .
fY (y )

NV HANH Probability and statistics First semester, 2023-2024 14 / 39

 Conditional probability distribution

Conditional probability distribution

Definitions 4.7:
Let (X , Y ) be a two-dimensional continuous random variable with the
joint pdf f (x, y ).
The conditional probability distribution of X givent that Y = y is
defined by conditional pdf:

f (x, y )
fX |Y =y (x) = .
fY (y )

The conditional probability distribution of Y givent that X = x is

defined by conditional pdf:

f (x, y )
fY |X =x (y ) = .
fX (x)

NV HANH Probability and statistics First semester, 2023-2024 14 / 39

 Independence

Independence

Definitions 4.8:
Let (X , Y ) be a two-dimensional discrete random variable with the
joint pmf f (x, y ). X and Y are called independent if and only if:

P(X = xi , Y = yj ) = P(X = xi )P(Y = yj )

for all i = 1, .., n; j = 1, ..., m.

NV HANH Probability and statistics First semester, 2023-2024 15 / 39

 Independence

Independence

Definitions 4.8:
Let (X , Y ) be a two-dimensional discrete random variable with the
joint pmf f (x, y ). X and Y are called independent if and only if:

P(X = xi , Y = yj ) = P(X = xi )P(Y = yj )

for all i = 1, .., n; j = 1, ..., m.

Let (X , Y ) be a two-dimensional continuous random variable with the
joint pdf f (x, y ). X and Y are called independent if and only if:

f (x, y ) = fX (x)fY (y )

for all (x, y ) ∈ R 2 .

NV HANH Probability and statistics First semester, 2023-2024 15 / 39

 Examples

Examples

Example 4.1:
Let X be the number of accidents per day and Y be the number of injured
or dead people per day at a small town. We observed X and Y during a
period of 100 days and obtained the following data:
X \Y 0 1 2 3
0 8 0 0 0
1 22 9 1 0
2 20 10 3 1
3 8 5 2 1
4 4 3 2 1

NV HANH Probability and statistics First semester, 2023-2024 16 / 39

 Examples

Examples

Example 4.1:
Let X be the number of accidents per day and Y be the number of injured
or dead people per day at a small town.
Find the joint probability distribution of (X , Y ).
Find the marginal probability distribution of X and of Y .
Find the conditional probability distribution of X given that Y = 2.
Find the conditional probability distribution of Y given that X = 2.
Are X and Y independent?

NV HANH Probability and statistics First semester, 2023-2024 17 / 39

 Examples

Examples

Solution of Example 4.1:

The joint probability distribution of (X , Y ) is the following table:
X \Y 0 1 2 3
0 0.08 0 0 0
1 0.22 0.09 0.01 0
2 0.2 0.1 0.03 0.01
3 0.08 0.05 0.02 0.01
4 0.04 0.03 0.02 0.01

NV HANH Probability and statistics First semester, 2023-2024 18 / 39

 Examples

Examples

Solution of Example 4.1:

To find the marginal probability distributions of X and of Y , we add the
probabilities by rows and by columns:
X \Y 0 1 2 3 P(xi )
0 0.08 0 0 0 0.08
1 0.22 0.09 0.01 0 0.32
2 0.2 0.1 0.03 0.01 0.34
3 0.08 0.05 0.02 0.01 0.16
4 0.04 0.03 0.02 0.01 0.1
P(yj ) 0.62 0.27 0.08 0.03 1

NV HANH Probability and statistics First semester, 2023-2024 19 / 39

 Examples

Examples

Solution of Example 4.1:

The marginal probability distribution of X is the following table:
X 0 1 2 3 4
P 0.08 0.32 0.34 0.16 0.1

NV HANH Probability and statistics First semester, 2023-2024 20 / 39

 Examples

Examples

Solution of Example 4.1:

The marginal probability distribution of X is the following table:
X 0 1 2 3 4
P 0.08 0.32 0.34 0.16 0.1
The marginal probability distribution of Y is the following table:
X 0 1 2 3
P 0.62 0.27 0.08 0.03

NV HANH Probability and statistics First semester, 2023-2024 20 / 39

 Examples

Examples

Solution of Example 4.1:

The conditional probability distribution of X given that Y = 2 is the
following table:
X |Y = 2 0 1 2 3 4
P 0 1/8 3/8 2/8 2/8

NV HANH Probability and statistics First semester, 2023-2024 21 / 39

 Examples

Examples

Solution of Example 4.1:

The conditional probability distribution of X given that Y = 2 is the
following table:
X |Y = 2 0 1 2 3 4
P 0 1/8 3/8 2/8 2/8
The conditional probability distribution of Y given that X = 2 is the
following table:
Y |X = 2 0 1 2 3
P 20/34 10/34 3/34 1/34

NV HANH Probability and statistics First semester, 2023-2024 21 / 39

 Examples

Examples

Solution of Example 4.1:

The conditional probability distribution of X given that Y = 2 is the
following table:
X |Y = 2 0 1 2 3 4
P 0 1/8 3/8 2/8 2/8
The conditional probability distribution of Y given that X = 2 is the
following table:
Y |X = 2 0 1 2 3
P 20/34 10/34 3/34 1/34
Since 0.08 = P(X = 0, Y = 0) 6= P(X = 0)P(Y = 0) =
0.08 ∗ 0.62 = 0.0496 then X and Y are not independent.

NV HANH Probability and statistics First semester, 2023-2024 21 / 39

 Examples

Examples

Example 4.2:
Suppose that the two-dimensional continuous random variable (X , Y ) has
the following joint pdf:
(
Cxy if 0 ≤ x ≤ 4; 1 ≤ y ≤ 5,
f (x, y ) =
0 otherwise

Find the normalizing constant C .

Find the marginal probability distribution of X and of Y .
Find the conditional probability distribution of X given that Y = 2.
Find the conditional probability distribution of Y given that X = 2.
Are X and Y independent?

NV HANH Probability and statistics First semester, 2023-2024 22 / 39

 Examples

Examples

Solution of Example 4.2:

Find the normalizing constant C .
Since f (x, y ) is a joint pdf then f (x, y ) ≥ 0, ∀(x, y ) ∈ R 2 ⇔ C ≥ 0.

NV HANH Probability and statistics First semester, 2023-2024 23 / 39

 Examples

Examples

Solution of Example 4.2:

Find the normalizing constant C .
Since f (x, y ) is a joint pdf then f (x, y ) ≥ 0, ∀(x, y ) ∈ R 2 ⇔ C ≥ 0.
RR R4R5 R4 R5 
And R 2 f (x, y )dxdy = 0 1 Cxydxdy = C 0 x 1 ydy dx =
96C = 1, so C = 1/96.

NV HANH Probability and statistics First semester, 2023-2024 23 / 39

 Examples

Examples

Solution of Example 4.2:

Find the normalizing constant C .
Since f (x, y ) is a joint pdf then f (x, y ) ≥ 0, ∀(x, y ) ∈ R 2 ⇔ C ≥ 0.
RR R4R5 R4 R5 
And R 2 f (x, y )dxdy = 0 1 Cxydxdy = C 0 x 1 ydy dx =
96C = 1, so C = 1/96.
Then (
1
if 0 ≤ x ≤ 4; 1 ≤ y ≤ 5,
96 xy
f (x, y ) =
0 otherwise

NV HANH Probability and statistics First semester, 2023-2024 23 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of X .
R +∞
The marginal pdf of X is fX (x) = −∞ f (x, y )dy .

NV HANH Probability and statistics First semester, 2023-2024 24 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of X .
R +∞
The marginal pdf of X is fX (x) = −∞ f (x, y )dy .
R +∞
If x ∈
/ [0, 4] then fX (x) = −∞ 0dy = 0.

NV HANH Probability and statistics First semester, 2023-2024 24 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of X .
R +∞
The marginal pdf of X is fX (x) = −∞ f (x, y )dy .
R +∞
If x ∈
/ [0, 4] then fX (x) = −∞ 0dy = 0.
R1 R5 1 R +∞
If x ∈ [0, 4] then fX (x) = −∞ 0dy + 1 96 xydy + 5 0dy = 18 x.

NV HANH Probability and statistics First semester, 2023-2024 24 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of X .
R +∞
The marginal pdf of X is fX (x) = −∞ f (x, y )dy .
R +∞
If x ∈
/ [0, 4] then fX (x) = −∞ 0dy = 0.
R1 R5 1 R +∞
If x ∈ [0, 4] then fX (x) = −∞ 0dy + 1 96 xydy + 5 0dy = 18 x.
Then (
1
8xif 0 ≤ x ≤ 4,
fX (x) =
0 otherwise

NV HANH Probability and statistics First semester, 2023-2024 24 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of Y .
R +∞
The marginal pdf of Y is fY (y ) = −∞ f (x, y )dx.

NV HANH Probability and statistics First semester, 2023-2024 25 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of Y .
R +∞
The marginal pdf of Y is fY (y ) = −∞ f (x, y )dx.
R +∞
If y ∈
/ [1, 5] then fY (y ) = −∞ 0dx = 0.

NV HANH Probability and statistics First semester, 2023-2024 25 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of Y .
R +∞
The marginal pdf of Y is fY (y ) = −∞ f (x, y )dx.
R +∞
If y ∈
/ [1, 5] then fY (y ) = −∞ 0dx = 0.
R0 R4 1 R +∞ 1
If y ∈ [1, 5] then fY (y ) = −∞ 0dx + 0 96 xydx + 4 0dx = 12 y .

NV HANH Probability and statistics First semester, 2023-2024 25 / 39

 Examples

Examples

Solution of Example 4.2:

Find the marginal probability distribution of Y .
R +∞
The marginal pdf of Y is fY (y ) = −∞ f (x, y )dx.
R +∞
If y ∈
/ [1, 5] then fY (y ) = −∞ 0dx = 0.
R0 R4 1 R +∞ 1
If y ∈ [1, 5] then fY (y ) = −∞ 0dx + 0 96 xydx + 4 0dx = 12 y .
Then (
1
if 1 ≤ y ≤ 5,
12 y
fY (y ) =
0 otherwise

NV HANH Probability and statistics First semester, 2023-2024 25 / 39

 Examples

Examples

Solution of Example 4.2:

Find the conditional probability distribution of X given that Y = 2.
The conditional pdf of X given that Y = 2 is
( (1/96)x.2
f (x, 2) 2/12 = 81 x if 0 ≤ x ≤ 4,
fX |Y =2 (x) = = 0
fY (2) 2/12 = 0 otherwise

NV HANH Probability and statistics First semester, 2023-2024 26 / 39

 Examples

Examples

Solution of Example 4.2:

Find the conditional probability distribution of Y given that X = 2.
The conditional pdf of Y given that X = 2 is
( (1/96)2y
1
f (2, y ) 2/8 = 12 y if 1 ≤ y ≤ 5,
fY |X =2 (y ) = = 0
fX (2) 2/8 = 0 otherwise

NV HANH Probability and statistics First semester, 2023-2024 27 / 39

 Examples

Examples

Solution of Example 4.2:

Find the conditional probability distribution of Y given that X = 2.
The conditional pdf of Y given that X = 2 is
( (1/96)2y
1
f (2, y ) 2/8 = 12 y if 1 ≤ y ≤ 5,
fY |X =2 (y ) = = 0
fX (2) 2/8 = 0 otherwise

Are X and Y independent?

Since f (x, y ) = fX (x)fY (y ), ∀(x, y ) ∈ R 2 then X and Y are
independent.

NV HANH Probability and statistics First semester, 2023-2024 27 / 39

 Covariance matrix and correlation

Functions of random vectors

Let (X , Y ) be a two-dimensional random variable and g : R 2 → R be a

function. The random variable Z is defined by Z = g (X , Y ). Then
E (Z ) = E [g (X , Y )] = ni=1 m
P P
j=1 g (xi , yj )pij if (X , Y ) is discrete.

NV HANH Probability and statistics First semester, 2023-2024 28 / 39

 Covariance matrix and correlation

Functions of random vectors

Let (X , Y ) be a two-dimensional random variable and g : R 2 → R be a

function. The random variable Z is defined by Z = g (X , Y ). Then
E (Z ) = E [g (X , Y )] = ni=1 m
P P
j=1 g (xi , yj )pij if (X , Y ) is discrete.
R +∞ R +∞
and E (Z ) = E [g (X , Y )] = −∞ −∞ g (x, y )f (x, y )dxdy if (X , Y ) is
continuous.

NV HANH Probability and statistics First semester, 2023-2024 28 / 39

 Covariance matrix and correlation

Covariance

Definitions 4.9:
Let (X , Y ) be a two-dimensional random variable. The covariance of
(X , Y ) is defined by:

cov (X , Y ) = E (XY ) − E (X )E (Y )

NV HANH Probability and statistics First semester, 2023-2024 29 / 39

 Covariance matrix and correlation

Covariance

Definitions 4.9:
Let (X , Y ) be a two-dimensional random variable. The covariance of
(X , Y ) is defined by:

cov (X , Y ) = E (XY ) − E (X )E (Y )

where
Pn Pn Pm
E (X ) = i=1 xi pi· = j=1 xi pij ,
Pm Pi=1n Pm
E (Y ) = j=1 yj p·j = i=1 j=1 yj pij ,
E (XY ) = ni=1 m
P P
j=1 xi yj pij ,
if (X , Y ) is discrete.

NV HANH Probability and statistics First semester, 2023-2024 29 / 39

 Covariance matrix and correlation

Covariance

Definitions 4.9:
Let (X , Y ) be a two-dimensional random variable. The covariance of
(X , Y ) is defined by:

cov (X , Y ) = E (XY ) − E (X )E (Y )

NV HANH Probability and statistics First semester, 2023-2024 30 / 39

 Covariance matrix and correlation

Covariance

Definitions 4.9:
Let (X , Y ) be a two-dimensional random variable. The covariance of
(X , Y ) is defined by:

cov (X , Y ) = E (XY ) − E (X )E (Y )

where
R +∞ R +∞ R +∞
E (X ) = −∞ xfX (x)dx = −∞ −∞ xf (x, y )dxdy ,
R +∞ R +∞ R +∞
E (Y ) = −∞ yfY (y )dy = −∞ −∞ yf (x, y )dxdy ,
R +∞ R +∞
E (XY ) = −∞ −∞ xyf (x, y )dxdy ,
if (X , Y ) is continuous.

NV HANH Probability and statistics First semester, 2023-2024 30 / 39

 Covariance matrix and correlation

Covariance

Properties:
If X and Y are independent then cov (X , Y ) = 0, but the opposite is
not always true.

NV HANH Probability and statistics First semester, 2023-2024 31 / 39

 Covariance matrix and correlation

Covariance

Properties:
If X and Y are independent then cov (X , Y ) = 0, but the opposite is
not always true.
If cov (X , Y ) 6= 0 then X and Y are called to be related and if
cov (X , Y ) = 0 then X and Y are not related.

NV HANH Probability and statistics First semester, 2023-2024 31 / 39

 Covariance matrix and correlation

Covariance matrix

Definitions 4.10:
Let (X , Y ) be a two-dimensional random variable. The covariance matrix
of (X , Y ) is defined by:
 
V (X ) cov (X , Y )
Γ(X , Y ) =
cov (Y , X ) V (Y )

where cov (Y , X ) = cov (X , Y ).

NV HANH Probability and statistics First semester, 2023-2024 32 / 39

 Covariance matrix and correlation

Coefficient of correlation
Definitions 4.11:
Let (X , Y ) be a two-dimensional random variable. The coefficient of
correlation between X and Y is defined by:

cov (X , Y )
ρ(X , Y ) = p .
V (X )V (Y )

NV HANH Probability and statistics First semester, 2023-2024 33 / 39

 Covariance matrix and correlation

Coefficient of correlation
Definitions 4.11:
Let (X , Y ) be a two-dimensional random variable. The coefficient of
correlation between X and Y is defined by:

cov (X , Y )
ρ(X , Y ) = p .
V (X )V (Y )

Properties:
If X and Y are independent then ρ(X , Y ) = 0, but the opposite is
not always true.

NV HANH Probability and statistics First semester, 2023-2024 33 / 39

 Covariance matrix and correlation

Coefficient of correlation
Definitions 4.11:
Let (X , Y ) be a two-dimensional random variable. The coefficient of
correlation between X and Y is defined by:

cov (X , Y )
ρ(X , Y ) = p .
V (X )V (Y )

Properties:
If X and Y are independent then ρ(X , Y ) = 0, but the opposite is
not always true.
If ρ(X , Y ) 6= 0 then X and Y are related and if ρ(X , Y ) = 0 then X
and Y are not related.

NV HANH Probability and statistics First semester, 2023-2024 33 / 39

 Covariance matrix and correlation

Coefficient of correlation
Definitions 4.11:
Let (X , Y ) be a two-dimensional random variable. The coefficient of
correlation between X and Y is defined by:

cov (X , Y )
ρ(X , Y ) = p .
V (X )V (Y )

Properties:
If X and Y are independent then ρ(X , Y ) = 0, but the opposite is
not always true.
If ρ(X , Y ) 6= 0 then X and Y are related and if ρ(X , Y ) = 0 then X
and Y are not related.
The range of ρ(X , Y ) is [−1, 1]: −1 ≤ ρ(X , Y ) ≤ 1.

NV HANH Probability and statistics First semester, 2023-2024 33 / 39

 Covariance matrix and correlation

Coefficient of correlation
Definitions 4.11:
Let (X , Y ) be a two-dimensional random variable. The coefficient of
correlation between X and Y is defined by:

cov (X , Y )
ρ(X , Y ) = p .
V (X )V (Y )

Properties:
If X and Y are independent then ρ(X , Y ) = 0, but the opposite is
not always true.
If ρ(X , Y ) 6= 0 then X and Y are related and if ρ(X , Y ) = 0 then X
and Y are not related.
The range of ρ(X , Y ) is [−1, 1]: −1 ≤ ρ(X , Y ) ≤ 1.
If Y = a + bX with b > 0 then ρ(X , Y ) = 1 and if Y = a + bX with
b < 0 then ρ(X , Y ) = −1.
NV HANH Probability and statistics First semester, 2023-2024 33 / 39
 Examples

Examples

Example 4.3:
Let X be the number of accidents per day and Y be the number of injured
or dead people per day at a small town. We observed X and Y during a
period of 100 days and obtained the following data:
X \Y 0 1 2 3
0 8 0 0 0
1 22 9 1 0
2 20 10 3 1
3 8 5 2 1
4 4 3 2 1
Find the covariance matrix of (X , Y ).
Find the coefficient of correlation between X and Y .
Are X and Y related?

NV HANH Probability and statistics First semester, 2023-2024 34 / 39

 Examples

Examples

Solution of Example 4.3:

Find the covariance matrix of (X , Y ).
E (XY ) = 0 ∗ 0 ∗ 0.08 + 1 ∗ 0 ∗ 0.22 + ... + 4 ∗ 3 ∗ 0.01 = 1.25

NV HANH Probability and statistics First semester, 2023-2024 35 / 39

 Examples

Examples

Solution of Example 4.3:

Find the covariance matrix of (X , Y ).
E (XY ) = 0 ∗ 0 ∗ 0.08 + 1 ∗ 0 ∗ 0.22 + ... + 4 ∗ 3 ∗ 0.01 = 1.25
Then
The covariance of (X , Y ) is
cov (X , Y ) = E (XY ) − E (X )E (Y ) = 0.2724

NV HANH Probability and statistics First semester, 2023-2024 35 / 39

 Examples

Examples

Solution of Example 4.3:

Find the covariance matrix of (X , Y ).
E (XY ) = 0 ∗ 0 ∗ 0.08 + 1 ∗ 0 ∗ 0.22 + ... + 4 ∗ 3 ∗ 0.01 = 1.25
Then
The covariance of (X , Y ) is
cov (X , Y ) = E (XY ) − E (X )E (Y ) = 0.2724
The covariance matrix of (X , Y ) is
   
V (X ) cov (X , Y ) 1.1856 0.2724
Γ(X , Y ) = =
cov (Y , X ) V (Y ) 0.2724 0.5896

NV HANH Probability and statistics First semester, 2023-2024 35 / 39

 Examples

Examples

Solution of Example 4.3:

Find the coefficient of correlation between X and Y .
The coefficient of correlation between X and Y is
cov (X , Y ) 0.2724
ρ(X , Y ) = p =√ = 0.823
V (X )V (Y ) 1.1856 ∗ 0.5896

NV HANH Probability and statistics First semester, 2023-2024 36 / 39

 Examples

Examples

Solution of Example 4.3:

Find the coefficient of correlation between X and Y .
The coefficient of correlation between X and Y is
cov (X , Y ) 0.2724
ρ(X , Y ) = p =√ = 0.823
V (X )V (Y ) 1.1856 ∗ 0.5896

Are X and Y related?

Since ρ(X , Y ) = 0.823 6= 0 then X and Y are related and since
ρ(X , Y ) = 0.823 is close to 1, then X and Y have a strong positive
linear relationship.

NV HANH Probability and statistics First semester, 2023-2024 36 / 39

 Examples

Examples

Example 4.4:
Suppose that the two-dimensional continuous random variable (X , Y ) has
the following joint pdf:
(
1
xy if 0 ≤ x ≤ 4; 1 ≤ y ≤ 5,
f (x, y ) = 96
0 otherwise

Find the covariance matrix of (X , Y ).

Find the coefficient of correlation between X and Y .
Are X and Y related?

NV HANH Probability and statistics First semester, 2023-2024 37 / 39

 Examples

Examples

Solution of Example 4.4:

We have
R +∞ R4 1 2 8
µX = E (X ) = −∞ xfX (x)dx = 0 8 x dx = 3

NV HANH Probability and statistics First semester, 2023-2024 38 / 39

 Examples

Examples

Solution of Example 4.4:

We have
R +∞ R4
µX = E (X ) = −∞ xfX (x)dx = 0 18 x 2 dx = 38
R +∞ R4
V (X ) = −∞ x 2 fX (x)dx − µ2X = 0 18 x 3 dx − (8/3)2 = 8
9

NV HANH Probability and statistics First semester, 2023-2024 38 / 39

 Examples

Examples

Solution of Example 4.4:

We have
R +∞ R4
µX = E (X ) = −∞ xfX (x)dx = 0 18 x 2 dx = 38
R +∞ R4
V (X ) = −∞ x 2 fX (x)dx − µ2X = 0 18 x 3 dx − (8/3)2 = 8
9
R +∞ R5 1 2
µY = E (Y ) = −∞ yfY (y )dy = 1 12 y dy = 31 9 .

NV HANH Probability and statistics First semester, 2023-2024 38 / 39

 Examples

Examples

Solution of Example 4.4:

We have
R +∞ R4
µX = E (X ) = −∞ xfX (x)dx = 0 18 x 2 dx = 38
R +∞ R4
V (X ) = −∞ x 2 fX (x)dx − µ2X = 0 18 x 3 dx − (8/3)2 = 98
R +∞ R5 1 2
µY = E (Y ) = −∞ yfY (y )dy = 1 12 y dy = 31 9 .
R +∞ 2 5 1 3 92
V (Y ) = −∞ y fY (y )dy − µ2Y = 1 12 y dy − (31/9)2 =
R
81 .

NV HANH Probability and statistics First semester, 2023-2024 38 / 39

 Examples

Examples

Solution of Example 4.4:

We have
R +∞ R4
µX = E (X ) = −∞ xfX (x)dx = 0 18 x 2 dx = 38
R +∞ R4
V (X ) = −∞ x 2 fX (x)dx − µ2X = 0 18 x 3 dx − (8/3)2 = 98
R +∞ R5 1 2
µY = E (Y ) = −∞ yfY (y )dy = 1 12 y dy = 31 9 .
R +∞ 2 5 1 3
V (Y ) = −∞ y fY (y )dy − µ2Y = 1 12 y dy − (31/9)2 = 92
R
81 .
R +∞ R +∞ 1
R 4 5 2 2
R 248
E (XY ) = −∞ −∞ xyf (x, y )dxdy = 96 0 1 x y dxdy = 27

NV HANH Probability and statistics First semester, 2023-2024 38 / 39

 Examples

Examples

Solution of Example 4.4:

Then
The covariance of (X , Y ) is
248 8 31
cov (X , Y ) = E (XY ) − E (X )E (Y ) = 27 − 3 9 = 0.

NV HANH Probability and statistics First semester, 2023-2024 39 / 39

 Examples

Examples

Solution of Example 4.4:

Then
The covariance of (X , Y ) is
248 8 31
cov (X , Y ) = E (XY ) − E (X )E (Y ) = 27 − 3 9 = 0.
The covariance matrix of (X , Y ) is
  8 
V (X ) cov (X , Y ) 0
Γ(X , Y ) = = 9 92
cov (Y , X ) V (Y ) 0 81

NV HANH Probability and statistics First semester, 2023-2024 39 / 39

 Examples

Examples

Solution of Example 4.4:

The coefficient of correlation between X and Y is
cov (X , Y ) 0
ρ(X , Y ) = p =p =0
V (X )V (Y ) (8/9)(92/81)

NV HANH Probability and statistics First semester, 2023-2024 40 / 39

 Examples

Examples

Solution of Example 4.4:

The coefficient of correlation between X and Y is
cov (X , Y ) 0
ρ(X , Y ) = p =p =0
V (X )V (Y ) (8/9)(92/81)

Are X and Y related?

Since ρ = 0 then X and Y are not related.

NV HANH Probability and statistics First semester, 2023-2024 40 / 39