Chapter 3: Continuous Random Variables

(1)
NGUYỄN THỊ THU THỦY

SCHOOL OF APPLIED MATHEMATICS AND INFORMATICS,

HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

HANOI, 10/2019

(1)
Email: thuy.nguyenthithu2@hust.edu.vn
Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 1 / 47
Introduction

Continuous Sample Space

This chapter analyzes random variables that range over continuous sets of
numbers.
A continuous set of numbers, sometimes referred to as an interval, contains all of
the real numbers between two limits.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 2 / 47
Introduction

Continuous Sample Space

This chapter analyzes random variables that range over continuous sets of
numbers.
A continuous set of numbers, sometimes referred to as an interval, contains all of
the real numbers between two limits.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 2 / 47
 3.1 The Cumulative Distribution Function 3.1.1 Cumulative Distribution Function

3.1.1 Cumulative Distribution Function

Definition 3.1 (Cumulative Distribution Function)

The cumulative distribution function (CDF) of random variable X is

FX (x) = P [X < x], x ∈ R. (1.1)

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 3 / 47
 3.1 The Cumulative Distribution Function 3.1.1 Cumulative Distribution Function

3.1.1 Cumulative Distribution Function

Theorem 3.1: For any random variable X

(a) FX (−∞) = 0.
(b) FX (+∞) = 1.
(c) P [a ≤ X < b] = FX (b) − FX (a).

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 4 / 47
 3.1 The Cumulative Distribution Function 3.1.2 Continuous Random Variable

3.1.2 Continuous Random Variable

Definition 3.2 (Continuous Random Variable)

X is a continuous random variable if the CDF FX (x) is a continuous function.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 5 / 47
 3.2 Probability Density Function 3.2.1 Definition

3.2.1 Definition

Definition 3.3 (Probability Density Function)

The probability density function (PDF) of a continuous random variable X is

dFX (x)
fX (x) = . (2.2)
dx

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 6 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Theorem 3.2: For a continuous random variable X with PDF fX (x)

(a) fX (x) ≥ 0 for all x,
+∞
Z
(b) fX (x)dx = 1,
−∞

Zx
(c) FX (x) = fX (u)du.
−∞

Theorem 3.3
Zb
P [a ≤ X < b] = fX (x)dx.
a

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 7 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Theorem 3.2: For a continuous random variable X with PDF fX (x)

(a) fX (x) ≥ 0 for all x,
+∞
Z
(b) fX (x)dx = 1,
−∞

Zx
(c) FX (x) = fX (u)du.
−∞

Theorem 3.3
Zb
P [a ≤ X < b] = fX (x)dx.
a

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 7 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties
Remark 3.4
(a) P [X = a] = 0 for continuous random variables. This implies that

P [X ≥ a] = P [X > a] and P [X ≤ a] = P [X < a].

This is not true in general for discrete random variables.

(b) The probability that X will fall into a particular interval say, from a to b is equal to
the area under the curve between the two points a and b.

Figure: P [a < X < b] is equal to the shaded area under the curve

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 8 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties
Remark 3.4
(a) P [X = a] = 0 for continuous random variables. This implies that

P [X ≥ a] = P [X > a] and P [X ≤ a] = P [X < a].

This is not true in general for discrete random variables.

(b) The probability that X will fall into a particular interval say, from a to b is equal to
the area under the curve between the two points a and b.

Figure: P [a < X < b] is equal to the shaded area under the curve

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 8 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties
Remark 3.4
(a) P [X = a] = 0 for continuous random variables. This implies that

P [X ≥ a] = P [X > a] and P [X ≤ a] = P [X < a].

This is not true in general for discrete random variables.

(b) The probability that X will fall into a particular interval say, from a to b is equal to
the area under the curve between the two points a and b.

Figure: P [a < X < b] is equal to the shaded area under the curve

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 8 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Example 3.4
For the CDF in (2.3), find the PDF of X and the probability of the event
{1/4 < X ≤ 3/4}.

0,
 x < 0,
FX (x) = x, 0 ≤ x < 1, (2.3)

1, x ≥ 1.


Example 3.4 Solution

The PDF of X is
(
1, 0 ≤ x < 1,
fX (x) = (2.4)
0, otherwise.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 9 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Example 3.4
For the CDF in (2.3), find the PDF of X and the probability of the event
{1/4 < X ≤ 3/4}.

0,
 x < 0,
FX (x) = x, 0 ≤ x < 1, (2.3)

1, x ≥ 1.


Example 3.4 Solution

The PDF of X is
(
1, 0 ≤ x < 1,
fX (x) = (2.4)
0, otherwise.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 9 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Example 3.4 Solution

To find the probability that X is between 1/4 and 3/4, we can use either Theorem 3.1
or Theorem 3.3. Thus

P [1/4 < X ≤ 3/4] = FX (3/4) − FX (1/4) = 1/2, (2.5)

and equivalently,

Z3/4 Z3/4
P [1/4 < X ≤ 3/4] = fX (x)dx = dx = 1/2. (2.6)
1/4 1/4

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 10 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Example 3.5
The CDF of Y is

0,
 y < 0,
FY (y) = y 3 , 0 ≤ y ≤ 1,

1, y > 1.


Find the PDF of Y and the probability that Y is between 1/4 and 3/4.

Example 3.5 Solution

Applying Definition 3.3,
(
3y 2 , 0 < y ≤ 1,
fY (y) =
0, otherwise.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 11 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Example 3.5
The CDF of Y is

0,
 y < 0,
FY (y) = y 3 , 0 ≤ y ≤ 1,

1, y > 1.


Find the PDF of Y and the probability that Y is between 1/4 and 3/4.

Example 3.5 Solution

Applying Definition 3.3,
(
3y 2 , 0 < y ≤ 1,
fY (y) =
0, otherwise.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 11 / 47
 3.2 Probability Density Function 3.2.2 Properties

3.2.2 Properties

Example 3.5 Solution (continuous)

Theorem 3.1 or Theorem 3.3 can be used to calculate the probability of observing Y
between 1/4 and 3/4:

P [1/4 < Y < 3/4] = FY (3/4) − FY (1/4) = (3/4)3 − (1/4)3 = 13/32,

and equivalently,

Z3/4 Z3/4
P [1/4 < Y < 3/4] = fY (y)dy = 3y 2 dy = 13/32.
1/4 1/4

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 12 / 47
 3.3 Expected Values 3.3.1 Expected Values

3.3.1 Expected Values

Definition 3.4 (Expected Value)

The expected value of a continuous random variable X is
+∞
Z
E[X] = xfX (x)dx. (3.7)
−∞

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 13 / 47
 3.3 Expected Values 3.3.1 Expected Values

3.3.1 Expected Values

Example 3.6
In Example 3.4, we found that the stopping point X of the spinning wheel experiment
was a uniform random variable with PDF
(
1, 0 ≤ x < 1,
fX (x) = (3.8)
0, otherwise.

Find the expected stopping point E[X] of the pointer.

Example 3.6 Solution

+∞
Z Z1
E[X] = xfX (x)dx = xdx = 1/2. (3.9)
−∞ 0

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 14 / 47
 3.3 Expected Values 3.3.1 Expected Values

3.3.1 Expected Values

Example 3.6
In Example 3.4, we found that the stopping point X of the spinning wheel experiment
was a uniform random variable with PDF
(
1, 0 ≤ x < 1,
fX (x) = (3.8)
0, otherwise.

Find the expected stopping point E[X] of the pointer.

Example 3.6 Solution

+∞
Z Z1
E[X] = xfX (x)dx = xdx = 1/2. (3.9)
−∞ 0

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 14 / 47
 3.3 Expected Values 3.3.1 Expected Values

3.3.1 Expected Values

Theorem 3.4: The expected value of a function, g(X), of random variable X is

+∞
Z
E[g(X)] = g(x)fX (x)dx.
−∞

Theorem 3.5: For any random variable X

(a) E[X − µX ] = 0, µX = E[X],
(b) E[aX + b] = aE[X] + b.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 15 / 47
 3.3 Expected Values 3.3.1 Expected Values

3.3.1 Expected Values

Theorem 3.4: The expected value of a function, g(X), of random variable X is

+∞
Z
E[g(X)] = g(x)fX (x)dx.
−∞

Theorem 3.5: For any random variable X

(a) E[X − µX ] = 0, µX = E[X],
(b) E[aX + b] = aE[X] + b.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 15 / 47
 3.3 Expected Values 3.3.2 Variance. Standard Deviation

3.3.2 Variance. Standard Deviation

Definition 3.5 (Variance. Standard Deviation)

The variance of a continuous random variable X is
+∞
Z
V ar[X] = (x − µX )2 fX (x)dx. (3.10)
−∞
p
The standard deviation of a continuous random variable X is σX = V ar[X].

Note
Theorem 3.4 states that E[X 2 ], the second moment of X, is the integral
+∞
Z
E[X 2 ] = x2 fX (x)dx. (3.11)
−∞

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 16 / 47
 3.3 Expected Values 3.3.2 Variance. Standard Deviation

3.3.2 Variance. Standard Deviation

Definition 3.5 (Variance. Standard Deviation)

The variance of a continuous random variable X is
+∞
Z
V ar[X] = (x − µX )2 fX (x)dx. (3.10)
−∞
p
The standard deviation of a continuous random variable X is σX = V ar[X].

Note
Theorem 3.4 states that E[X 2 ], the second moment of X, is the integral
+∞
Z
E[X 2 ] = x2 fX (x)dx. (3.11)
−∞

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 16 / 47
 3.3 Expected Values 3.3.2 Variance. Standard Deviation

3.3.2 Variance. Standard deviation

Theorem 3.6
For any random variable X,
(a) V ar[X] = E[X 2 ] − µ2X , µX = E[X],
(b) V ar[aX + b] = a2 V ar[X].

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 17 / 47
 3.3 Expected Values 3.3.2 Variance. Standard Deviation

3.3.2 Variance. Standard deviation

Example 3.8
Find the variance and standard deviation of the pointer position in Example 3.1.

Example 3.8 Solution

To compute V ar[X], we use Theorem 3.5(c): V ar[X] = E[X 2 ] − µ2X . We calculate
E[X 2 ] directly from Theorem 3.4 with g(X) = X 2 :
+∞
Z Z1
2
E[X ] = 2
x fX (x)dx = x2 dx = 1/3.
−∞ 0

In Example 3.6, we have E[X]p= 1/2. Thus V√ar[X] = 1/3 − (1/2)2 = 1/12, and the
standard deviation is σ[X] = V ar[X] = 1/ 12 = 0.289 meters.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 18 / 47
 3.3 Expected Values 3.3.2 Variance. Standard Deviation

3.3.2 Variance. Standard deviation

Example 3.8
Find the variance and standard deviation of the pointer position in Example 3.1.

Example 3.8 Solution

To compute V ar[X], we use Theorem 3.5(c): V ar[X] = E[X 2 ] − µ2X . We calculate
E[X 2 ] directly from Theorem 3.4 with g(X) = X 2 :
+∞
Z Z1
2
E[X ] = 2
x fX (x)dx = x2 dx = 1/3.
−∞ 0

In Example 3.6, we have E[X]p= 1/2. Thus V√ar[X] = 1/3 − (1/2)2 = 1/12, and the
standard deviation is σ[X] = V ar[X] = 1/ 12 = 0.289 meters.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 18 / 47
 3.3 Expected Values 3.3.2 Variance. Standard Deviation

3.3.2 Variance. Standard deviation

Example 3.9
Find the variance and standard deviation of Y , the maximum pointer position after
three spins, in Example 3.5.

Example 3.9 Solution

We proceed as in Example 3.8. We have fY (y) from Example 3.5 and E[Y ] = 3/4 from
Example 3.7:
+∞
Z Z1
E[Y 2 ] = y 2 fY (y)dy = y 2 (3y 2 )dy = 3/5.
−∞ 0

Thus the variance is

V ar[Y ] = 3/5 − (3/4)2 = 3/80 m2 ,

and the standard deviation is σ[Y ] = 0.194 meters.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 19 / 47
 3.3 Expected Values 3.3.2 Variance. Standard Deviation

3.3.2 Variance. Standard deviation

Example 3.9
Find the variance and standard deviation of Y , the maximum pointer position after
three spins, in Example 3.5.

Example 3.9 Solution

We proceed as in Example 3.8. We have fY (y) from Example 3.5 and E[Y ] = 3/4 from
Example 3.7:
+∞
Z Z1
E[Y 2 ] = y 2 fY (y)dy = y 2 (3y 2 )dy = 3/5.
−∞ 0

Thus the variance is

V ar[Y ] = 3/5 − (3/4)2 = 3/80 m2 ,

and the standard deviation is σ[Y ] = 0.194 meters.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 19 / 47
 3.3 Expected Values 3.3.3 Mode. Median

3.3.3 Mode. Median

Example 3.10
The probability density function of the continuous random variable X is

 3 x(2 − x), 0 ≤ x ≤ 2,
4
0, otherwise.

What is xmod ? What is xmed ?

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 20 / 47
 3.3 Expected Values 3.3.3 Mode. Median

3.3.3 Mode. Median

Example 3.10 Solution

Applying Theorem 3.2(c),


0, x ≤ 0,
  3
FX (x) = 3 x2 − x , 0 < x ≤ 2,
4 3


1, x > 2.

1
So xmed is a solution of the equation FX (x) = , or x3 − 3x2 + 2 = 0 with 0 < x ≤ 2.
2
Hence xmed = 1. 

 0, x ≤ 0,
3

0
Taking the derivative of the PDF fX (x), g(x) := fX (x) = (1 − x), 0 < x < 2,

 2
0, otherwise.

We can see the function g(x) reaches maximum at x = 1, so xmod = 1.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 21 / 47
 3.3 Expected Values 3.3.3 Mode. Median

3.4.2 Exponential Random Variable

Problem 3.1
Let X be a random variable denoting the proportion of students in a class who get a
grade lower than C. Suppose X is a random variable with the following probability
density function: (
k2 x(4 − x), 0 ≤ x ≤ 4,
fX (x) =
0, otherwise.

(a) Find k such that fX (x) is a probability density function.

(b) Find E[X] and V ar[X].
(c) Find the cumulative distribution function of X.
(d) A class is assumed to be unsuccessful if the proportion of students with a grade
lower than C is greater than 0.5. What is the probability that a class is
unsuccessful?
(e) Assume 5 classes take the course and their successes are independent of each
other. What is the probability that at least two of the classes will be unsuccessful?

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 22 / 47
 3.4 Families of Continuous Random Variables 3.4.1 Uniform Random Variable

3.4.1 Uniform Random Variable

Definition 3.5 (Uniform Random Variable)

X is a uniform (a, b) random variable if the PDF of X is

 1 , a ≤ x < b,
fX (x) = b − a (4.12)
0, otherwise,

where the two parameters are b > a.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 23 / 47
 3.4 Families of Continuous Random Variables 3.4.1 Uniform Random Variable

3.4.1 Uniform Random Variable

Definition 3.5 (Uniform Random Variable)

X is a uniform (a, b) random variable if the PDF of X is

 1 , a ≤ x < b,
fX (x) = b − a (4.12)
0, otherwise,

where the two parameters are b > a.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 23 / 47
 3.4 Families of Continuous Random Variables 3.4.1 Uniform Random Variable

3.4.1 Uniform Random Variable

Theorem 3.6
If X is a uniform (a, b) random variable,
(a) The CDF of X is


0, x ≤ a,
x − a
FX (x) = , a < x ≤ b, (4.13)

 b−a
1, x > b.


a+b
(b) The expected value of X is E[X] = .
2
2
(b − a)
(c) The variance of X is V ar[X] = .
12

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 24 / 47
 3.4 Families of Continuous Random Variables 3.4.1 Uniform Random Variable

3.4.1 Uniform Random Variable

Example 3.11
The phase angle, ϕ, of the signal at the input to a modem is uniformly distributed
between 0 and 2π radians. Find the CDF, the expected value, and the variance of ϕ.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 25 / 47
 3.4 Families of Continuous Random Variables 3.4.1 Uniform Random Variable

3.4.1 Uniform Random Variable

Example 3.11 Solution

From the problem statement, we identify the parameters of the uniform (a, b) random
variable as a = 0 and b = 2π. Therefore the PDF of is

 1 , 0 ≤ x < 2π,
fϕ (x) = 2π
0, otherwise.

The CDF is

0,x

 x ≤ 0,
Fϕ (x) = , 0 < x ≤ 2π,

 2π
1, x > 2π.

The expected value is E[ϕ] = b/2 = π radians, and the variance is

V ar[ϕ] = (2π)2 /12 = π 2 /3 rad2 .

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 26 / 47
 3.4 Families of Continuous Random Variables 3.4.2 Exponential Random Variable

3.4.2 Exponential Random Variable

Definition 3.6 (Exponential Random Variable)

X is an exponential (λ) random variable if the PDF of X is
(
λe−λx , x ≥ 0,
fX (x) = (4.14)
0, otherwise,

where the parameter λ > 0.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 27 / 47
 3.4 Families of Continuous Random Variables 3.4.2 Exponential Random Variable

3.4.2 Exponential Random Variable

Example 3.12
The probability that a telephone call lasts no more than t minutes is often modeled as
an exponential CDF.
(
1 − e−t/3 , t ≥ 0,
FT (t) = (4.15)
0, otherwise.

What is the PDF of the duration in minutes of a telephone conversation? What is the
probability that a conversation will last between 2 and 4 minutes?

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 28 / 47
 3.4 Families of Continuous Random Variables 3.4.2 Exponential Random Variable

3.4.2 Exponential Random Variable

Example 3.12 Solution

We find the PDF of T by taking the derivative of the CDF:

 1 e−t/3 , t ≥ 0,
dFT (t)
fT (t) = = 3 (4.16)
dt 0, otherwise.

Therefore, observing Definition 3.6, we recognize that T is an exponential (λ = 1/3)

random variable. The probability that a call lasts between 2 and 4 minutes is

P [2 ≤ T ≤ 4] = FT (4) − FT (2) = e−2/3 − e−4/3 = 0.250. (4.17)

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 29 / 47
 3.4 Families of Continuous Random Variables 3.4.2 Exponential Random Variable

3.4.2 Exponential Random Variable

Theroem 3.9
If X is an exponential (λ) random variable,
(
1 − e−λx , x ≥ 0,
(a) FX (x) =
0, otherwise.
(b) E[X] = 1/λ.
(c) V ar[X] = 1/λ2 .

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 30 / 47
 3.4 Families of Continuous Random Variables Problems

3.4.2 Exponential Random Variable

Problem 3.2
Let X be an exponential random variable with parameter λ and define Y = [X], the
largest integer in X, (ie. [x] = 0 for 0 ≤ x < 1, [x] = 1 for 1 ≤ x < 2 etc.)
(a) Find the probability mass function for Y .
(b) Find E(Y ).
(c) Find the cumulative distribution function of Y .
(d) Let Y represent the number of periods that a machine is in use before failure.
What is the probability that the machine is still working at the end of 10th period
given that it does not fail before 6th period?

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 31 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Definition 3.8 (Gaussian Random Variable)

X is a Gaussian (µ, σ) random variable if the PDF of X is

(x − µ)2
1 −
fX (x) = √ e 2σ 2 , (4.18)
σ 2π
where the parameter µ can be any real number and the parameter σ > 0.

Note
Many statistics texts use the notation X is N [µ, σ 2 ] as shorthand for X is a Gaussian
(µ, σ) random variable. In this notation, the N denotes normal. The graph of fX (x) has
a bell shape, where the center of the bell is x = µ and σ reflects the width of the bell. If
σ is small, the bell is narrow, with a high, pointy peak.
√ If σ is large, the bell is wide,
with a low, flat peak. (The height of the peak is 1/σ 2π).

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 32 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Definition 3.8 (Gaussian Random Variable)

X is a Gaussian (µ, σ) random variable if the PDF of X is

(x − µ)2
1 −
fX (x) = √ e 2σ 2 , (4.18)
σ 2π
where the parameter µ can be any real number and the parameter σ > 0.

Note
Many statistics texts use the notation X is N [µ, σ 2 ] as shorthand for X is a Gaussian
(µ, σ) random variable. In this notation, the N denotes normal. The graph of fX (x) has
a bell shape, where the center of the bell is x = µ and σ reflects the width of the bell. If
σ is small, the bell is narrow, with a high, pointy peak.
√ If σ is large, the bell is wide,
with a low, flat peak. (The height of the peak is 1/σ 2π).

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 32 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Figure: Normal probability distribution

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 33 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Figure: Normal probability distributions with differing values of µ and σ

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 34 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Figure: Two examples of a Gaussian random variable X with expected value µ

and standard deviation σ

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 35 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Theorem 3.11
If X is a Gaussian (µ, σ) random variable,

E[X] = µ, V ar[X] = σ 2 .

Theorem 3.12
If X is Gaussian (µ, σ), Y = aX + b is Gaussian (aµ + b, aσ).

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 36 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Theorem 3.11
If X is a Gaussian (µ, σ) random variable,

E[X] = µ, V ar[X] = σ 2 .

Theorem 3.12
If X is Gaussian (µ, σ), Y = aX + b is Gaussian (aµ + b, aσ).

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 36 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Definition 3.9 (Standard Normal Random Variable)

The standard normal random variable Z is the Gaussian (0, 1) random variable.

Note
Theorem 3.11 indicates that E[Z] = 0 and V ar[Z] = 1. The tables that we use to find
integrals of Gaussian PDFs contain values of FZ (z), the CDF of Z. We introduce the
special notation Φ(z) for this function.

Definition 3.10 (Standard Normal CDF)

The CDF of the standard normal random variable Z is
Zz
1 u2
Φ(z) = √ e− 2 du.
2π
−∞

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 37 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Definition 3.9 (Standard Normal Random Variable)

The standard normal random variable Z is the Gaussian (0, 1) random variable.

Note
Theorem 3.11 indicates that E[Z] = 0 and V ar[Z] = 1. The tables that we use to find
integrals of Gaussian PDFs contain values of FZ (z), the CDF of Z. We introduce the
special notation Φ(z) for this function.

Definition 3.10 (Standard Normal CDF)

The CDF of the standard normal random variable Z is
Zz
1 u2
Φ(z) = √ e− 2 du.
2π
−∞

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 37 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Definition 3.9 (Standard Normal Random Variable)

The standard normal random variable Z is the Gaussian (0, 1) random variable.

Note
Theorem 3.11 indicates that E[Z] = 0 and V ar[Z] = 1. The tables that we use to find
integrals of Gaussian PDFs contain values of FZ (z), the CDF of Z. We introduce the
special notation Φ(z) for this function.

Definition 3.10 (Standard Normal CDF)

The CDF of the standard normal random variable Z is
Zz
1 u2
Φ(z) = √ e− 2 du.
2π
−∞

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 37 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Theorem 3.13
If X is a Gaussian (µ, σ) random variable, the CDF of X is
x − µ
FX (x) = Φ .
σ
The probability that X is in the interval (a, b] is
b − µ a − µ
P [a ≤ X < b] = Φ −Φ .
σ σ

Note
In using this theorem, we transform values of a Gaussian random variable, X, to
equivalent values of the standard normal random variable, Z. For a sample value x of
the random variable X, the corresponding sample value of Z is
x−µ
z= . (4.19)
σ

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 38 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.15
Suppose your score on a test is x = 46, a sample value of the Gaussian (61, 10) random
variable. Express your test score as a sample value of the standard normal random
variable, Z.

Example 3.15 Solution

Equation (4.19) indicates that z = (46 − 61)/10 = −1.5. Therefore your score is 1.5
standard deviations less than the expected value.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 39 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.15
Suppose your score on a test is x = 46, a sample value of the Gaussian (61, 10) random
variable. Express your test score as a sample value of the standard normal random
variable, Z.

Example 3.15 Solution

Equation (4.19) indicates that z = (46 − 61)/10 = −1.5. Therefore your score is 1.5
standard deviations less than the expected value.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 39 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Note
To find probabilities of Gaussian random variables, we use the values of z presented in
Table 3.1. Note that this table contains entries only for z ≥ 0. For negative values of z,
we apply the following property of Φ(z).

Theorem 3.14

Φ(−z) = 1 − Φ(z).

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 40 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Note
To find probabilities of Gaussian random variables, we use the values of z presented in
Table 3.1. Note that this table contains entries only for z ≥ 0. For negative values of z,
we apply the following property of Φ(z).

Theorem 3.14

Φ(−z) = 1 − Φ(z).

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 40 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.17
If X is the Gaussian (61, 10) random variable, what is P [X ≤ 46]?

Example 3.17 Solution

Applying Theorem 3.13, Theorem 3.14 and the result of Example 3.15, we have

P [X ≤ 46] = Φ(−1.5) = 1 − Φ(1.5) = 1 − 0.93319 = 0.06681. (4.20)

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 41 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.17
If X is the Gaussian (61, 10) random variable, what is P [X ≤ 46]?

Example 3.17 Solution

Applying Theorem 3.13, Theorem 3.14 and the result of Example 3.15, we have

P [X ≤ 46] = Φ(−1.5) = 1 − Φ(1.5) = 1 − 0.93319 = 0.06681. (4.20)

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 41 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.18
If X is a Gaussian random variable with µ = 61 and σ = 10, what is P [51 < X ≤ 71]?

Example 3.17 Solution

Applying Equation (4.19), we find that the event {51 < X ≤ 71} corresponds to
{−1 < Z ≤ 1}. The probability of this event is

Φ(1) − Φ(−1) = Φ(1) − [1 − Φ(1)] = 2Φ(1) − 1 = 0.68268.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 42 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.18
If X is a Gaussian random variable with µ = 61 and σ = 10, what is P [51 < X ≤ 71]?

Example 3.17 Solution

Applying Equation (4.19), we find that the event {51 < X ≤ 71} corresponds to
{−1 < Z ≤ 1}. The probability of this event is

Φ(1) − Φ(−1) = Φ(1) − [1 − Φ(1)] = 2Φ(1) − 1 = 0.68268.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 42 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Remark 3.12 (The Normal Approximation to the Binomial Probability Distribution)

Let X be a binomial random variable with n trials and probability p of success. The
probability distribution of X is approximated using a normal curve with
√
µ = np and σ = npq,

and
 b + 0.5 − µ   a − 0.5 − µ 
P [a < X < b] = Φ −Φ . (4.21)
σ σ
This approximation is adequate as long as n is large and p is not too close to 0 or 1.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 43 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.20
Use the normal curve to approximate the probability that X = 8, 9, or 10 for a binomial
random variable with n = 25 and p = 0.5. Compare this approximation to the exact
binomial probability.

Example 3.20 Solution

You can find the exact binomial probability for this example because there are
cumulative binomial tables for n = 25,
8 9 10

P [X = 8] + P [X = 9] + P [X = 10] = C25 + C25 + C25 (0.5)25 ' 0.190535.

To use the normal approximation, first find the appropriate mean and standard deviation
√
for the normal curve: µ = np = 12.5, σ = npq = 2.5. It follows from (4.21) that

P [8 ≤ X ≤ 10] = Φ(−0.8) − Φ(−2) = 0.21186 − 0.02275 = 0.18911.

You can compare the approximation, 0.18911, to the actual probability, 0.190535. They
are quite close!

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 44 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.20
Use the normal curve to approximate the probability that X = 8, 9, or 10 for a binomial
random variable with n = 25 and p = 0.5. Compare this approximation to the exact
binomial probability.

Example 3.20 Solution

You can find the exact binomial probability for this example because there are
cumulative binomial tables for n = 25,
8 9 10

P [X = 8] + P [X = 9] + P [X = 10] = C25 + C25 + C25 (0.5)25 ' 0.190535.

To use the normal approximation, first find the appropriate mean and standard deviation
√
for the normal curve: µ = np = 12.5, σ = npq = 2.5. It follows from (4.21) that

P [8 ≤ X ≤ 10] = Φ(−0.8) − Φ(−2) = 0.21186 − 0.02275 = 0.18911.

You can compare the approximation, 0.18911, to the actual probability, 0.190535. They
are quite close!

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 44 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Note
The normal approximation to the binomial probabilities will be adequate if both np > 5
and n(1 − p) > 5.

Example 3.21
The reliability of an electrical fuse is the probability that a fuse, chosen at random from
production, will function under its designed conditions. A random sample of 1000 fuses
was tested and X = 27 defectives were observed. Calculate the approximate probability
of observing 27 or more defectives, assuming that the fuse reliability is 0.98.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 45 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Note
The normal approximation to the binomial probabilities will be adequate if both np > 5
and n(1 − p) > 5.

Example 3.21
The reliability of an electrical fuse is the probability that a fuse, chosen at random from
production, will function under its designed conditions. A random sample of 1000 fuses
was tested and X = 27 defectives were observed. Calculate the approximate probability
of observing 27 or more defectives, assuming that the fuse reliability is 0.98.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 45 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.21 Solution

The probability of observing a defective when a single fuse is tested is p = 0.02, given
√
that the fuse reliability is 0.98. Then µ = np = 20, σ = npq = 4.43.
The probability of 27 or more defective fuses, given n = 1000, is

P [X ≥ 27] = P [X = 27] + P [X = 28] + · · · + P [X = 1000].

It is appropriate to use the normal approximation to the binomial probability because

np = 20 and nq = 980 are both greater than 5. So
 1000 + 0.5 − 20   27 − 0.5 − 20 
P [27 ≤ X ≤ 1000] = Φ −Φ
4.43 4.43
= 1 − Φ(1.47) = 1 − 0.92922 = 0.07078.

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 46 / 47
 3.4 Families of Continuous Random Variables 3.4.3 Gaussian Random Variables

3.4.3 Gaussian Random Variables

Example 3.21 Solution

Figure: Normal approximation to the binomial for Example 3.21

Nguyễn Thị Thu Thủy (SAMI-HUST) Continuous Random Variables HANOI, 10/2019 47 / 47