Chapter 4

Continuous Random Variables and

Probability Distributions

Chaoyue Liu
Department of Mathematics and Statistics

1
Outline
• Continuous Random Variables
• PDFs and CDFs of Continuous Random Variables
• Expected Value and Variance of Continuous Random
Variables

• Uniform Distribution
• Exponential Distribution

• Normal Distribution, Standardization and Z-table

2
Introduction to Continuous Random
Variables

Chaoyue Liu
Department of Mathematics and Statistics

3
Outline
• Introduction to Continuous Random Variables
• Probability Density Function (pdf)
• Cumulative Distribution Function (cdf)
• Expected Values and Variance

4
Continuous Random Variables
• Discrete Random Variables: countable values
• Continuous Random Variables: uncountable values

• A random variable X is continuous if:

- its set of possible values is an entire interval of numbers;
• Examples:
- The lifetime of a product
- The waiting time
- other measured data: length, height, weight etc…

5
Probability Density Function (pdf)
• Discrete: probability mass function (pmf)
• Continuous: probability density function (pdf)

• Let X be a continuous random variable. The probability density function (pdf) of X is a real valued function f (x) that
satisfies

- f (x) ≥ 0 for any x ∈ ℝ

b

∫a
For any two real numbers a ≤ b ∈ ℝ, P(a ≤ X ≤ b) = f (x)dx
-

∫−∞
-
f (x)dx = 1

Example:

6
Probability Density Function (pdf)
• pdf f(x) is not a probability
• The probability for a continuous random variable is given by areas under pdf,
b

∫a
P(a ≤ X ≤ b) = f(x)dx

• We only talk about the probability of a continuous rv taking the value in an

interval, not at a point.

• P(X = c) = 0 for any number c ∈ ℝ.

• Equal sign doesn’t matter to continuous rv:
b

∫a
P(a ≤ X ≤ b) = P(a < X < b) = P(a ≤ X < b) = P(a < X ≤ b) = f(x)dx

7
Cumulative Distribution Function
• Let X have pdf f(x), then the cdf F(x) is given by
x

∫−∞
F(x) = P(X ≤ x) = f(t)dt, for x ∈ℝ

- For x ∈ ℝ, F(x) is the area under the density curve to the left of x.
- F(X) is non-decreasing

- lim F(x) = 0 and lim F(x) = 1

x→−∞ x→∞

• Example:
pdf cdf

8
Relationship between PDF and CDF for
Continuous Random Variables
• cdf can be found by integrating the pdf:
x

∫−∞
F(x) = f(t)dt

• pdf can be found by differentiating the cdf:

d
f(x) = F′(x) = F(x)
dx
b

∫a
Compute probability: P(a ≤ X ≤ b) = f(x)dx = F(b) − F(a)
•

9
Example (pdf -> cdf)
• Given the pdf of X as below,

{ 0,
kx 2, 0 ≤ x ≤ 2
f(x) =
otherwise

1. Find the value of k

2. Find the cdf of X

3. What is the probability that X>1?

n+1
x
∫
(hint: x ndx = + c)
n+1

10
Example (cdf -> pdf)
• Let X be a random variable with a cdf:
0, x≤0

[ ]
1 + ln ( x ) , 0 < x ≤ 4
x 4
F(x) = 4

1, 4<x

Find:

- (1) P(X ≤ 1)
- (2) P(1 ≤ X ≤ 3)
- (3) the pdf of X

11
Percentiles
• A percentile is a value below which a percentage of data falls. e.g. A
student’s test score is at the 85th percentile of the class means that 85% students
scores are lower than that score and 15% are above.

• Let p be a number between 0 and 1. The (100p)th percentile of the

distribution of a continuous rv X denoted by η(p), is defined by
η(p)

∫−∞
p = P(X ≤ η(p)) = F(η(p)) = f(u)du

• Special cases:

- Median ↔ 50th percentile.

- 1st Quantile ↔ 25th percentile.

- 3rd Quantile ↔ 75th percentile.

12
Example
• Let X be a continuous rv with the following pdf:

f(x) = 2x, 0 ≤ x ≤ 1
- Find

- (1) the 100p thpercentile of X

- (2) the median of X

13
Expected Value and Variance of
Continuous Random Variables
• The expected value or mean of a continuous rv X with pdf f(x) is
∞

∫−∞
μ = E(X ) = xf(x)dx

• Let h(x) be a function of X, then

∞

∫−∞
μh(X) = E(h(X )) = h(u)f(u)du

• The variance of X is
∞

∫−∞
V(X ) = σX2 = σ 2 = E [(X − μ)2] = (x − μ)2 f(x)dx

• The standard deviation of X is σ = σ2

• Short-cut formula: σ
2
= E[(X − μ)2] = E (X 2) − μ 2

• Linear function of X:

- Mean: E[aX + b] = aE[X] + b Variance: V(aX + b) = a 2V(X )

14
Example
• Let X be a random variable with the following pdf
1

{0
x ∈ [0,1] ∪ [2,3]
f(x) = 2
•
otherwise

- (1) Check that f(x) is a legitimate pdf

- (2) Compute E(X) and V(X)

- (3) Compute P(0.5 ≤ X ≤ 1.5)

15
Discrete vs Continuous

Discrete Random Variables Continuous Random Variables

pmf : P(x) = P{X = x} pdf : f(x) = F′(x)

∑ ∫A
P{X ∈ A} = p(x) P{X ∈ A} = f(x)dx
x∈A

x
∑ ∫−∞
F(x) = P{X ≤ x} = p(y) F(x) = P{X ≤ x} = f(y)dy
y≤x
∞

∑ ∫−∞
p(x) = 1 f(x)dx = 1
x

∞
∑
∫−∞
E(X ) = xp(x) E(X) = xf(x)dx
x
∞

∫−∞
(x − μ)2 p(x) V(X) = (x − μ)2 f(x)dx
∑
V(X) =
x

16
Uniform Distribution and Exponential
Distribution

Chaoyue Liu
Department of Mathematics and Statistics

17
Outline
• Uniform Distribution

• Exponential Distribution

18
Uniform Distribution
• A continuous random variable X is said to have a uniform distribution on the
interval [a, b], denoted by X ∼ uniform(a, b), if its pdf is
1

{ 0,
, a≤x≤b
f(x; a, b) = b−a
otherwise

b+a
- Mean: E(X) =
2
(b − a)2
- Variance: V(X) =
12
- The uniform distribution assigns equal probabilities to intervals of equal lengths

- For any real numbers c1, c2

∈ ℝ such that a ≤ c1 ≤ c2 ≤ b,
c2 − c1
P(c1 ≤ X ≤ c2) =
b−a

19
Exercise
• Check that f(x) for Uniform Distribution is a legitimate pdf.

• Derive the cdf of Uniform Distribution.

20
Exponential Distribution
• A random variable X follows an exponential distribution with (scale)
parameter λ > 0 , denoted by X ∼ Exp(λ), if the pdf of X is

{ 0
−λx
f(x) = λe x≥0
• otherwise

• If the number of Independent events occurring follows a Poisson distribution with rate
λ, then the distribution of waiting time between two consecutive events follows
exponential distribution with same parameter λ.

21
Properties of Exponential Distribution
• If X ∼ Exp(λ), then
1
- Mean: E(X) =
λ
1
- Variance: V(X) = 2
λ

{1 − e −λx x ≥ 0
0 x<0
cdf of X: F(x) =
-

- Memoryless Property, i.e. P(X > t + s | X > s) = P(X > t), for any
t, s ≥ 0. i.e. It "forgets" the time already spent.

22
Example
• Assume that buses arrive at a bus stop with rate α = 4 buses/hour. You
get off a bus and wait for the next bus to arrive.
- (1) What is the probability you wait between 15 to 30 minutes?
- (2) What is the 40th percentile of your waiting time
- (3) What is your expected waiting time?

23
Normal Distribution, Standardization
and Z-table

Chaoyue Liu
Department of Mathematics and Statistics

24
Outline
• Normal Distribution
• Standardization
• Z-table

25
Normal Distribution
• Normal distribution, also known as Gaussian distribution, is very important in
statistics, because it occurs naturally in many situations. e.g. height of the
population, IQ, measurement error…

• Normal distribution has a symmetric bell-shaped curve (most of values tend to

cluster around the mean, and the further a value is from the mean, the less likely it
is to occur)

26
Normal Distribution
• X follows a normal distribution with mean μ and variance σ 2 denoted by X ∼ N(μ, σ 2), if the pdf of
X is
1 − (x − μ)
2

f(x) = e 2σ 2 , −∞<x<∞
2πσ

where μ is a real number and σ 2 is a positive number

• N(μ, σ 2) is symmetric at μ
• Mean: E(X ) = μ, is a location parameter.
• Variance: V(X ) = σ 2, is a shape parameter. The larger the σ, the more the spread.

27
Standard Normal Distribution
• The normal distribution with μ = 0 and σ = 1 is called a standard normal
distribution.

2 X−μ
• If X ∼ N(μ, σ ), then the random variable Z = will follow the standard
σ
normal distribution Z ∼ N(0,1). Z is usually called Z value or Z score.

• The cdf of Z ∼ N(0,1), usually denoted by Φ(z) = P(Z ≤ z).

• Critical value zα , is a value such that P(Z ≥ zα) = α. zα is also the
100(1 − α)th percentile

e.g.
P(Z ≥ z0.05) = 0.05 ⟺ P(Z ≤ z0.05) = 1 − 0.05 = 0.95

so Z0.05 is also the 95th percentile of Z

28
Standardization
• Standardization:
2 X−μ
If X ∼ N(μ, σ ), then the random variable Z = ∼ N(0,1) , and the probability
σ
of X can be computed as

( σ σ )
a−μ b−μ
P(a ≤ X ≤ b) = P ≤Z≤

( σ ) ( σ )
b−μ a−μ
=Φ −Φ

• compute the cdf of normal distribution is hard in general

• any normal distribution can be converted to standard normal distribution
• one standard normal table (Z table) is only needed to find probabilities for
any normal random variables.

29
 Z table
• Z tables are composed as follows:
1. The row label contains the integer part and the first decimal place of
Z.
2. The column lable contains the second decimal place of Z
3. The values within the table are the probabilities

Example:

P(Z ≤ −2.6 5 ) = 0.004

⏟⏟ ⏟
1 2 3

30
Examples
• given z value, find probability

e.g. what is the probability P(Z ≤ − 2.52) ?

• given probability, find z value

e.g. what is the 80th percentile of Z ?

⟺ Φ(z) = 0.8, what is the value of z ?

31
Examples
• Some useful rules for Z ∼ N(0,1)
- P(Z > a) = 1 − Φ(a) - P( | Z | > a) = 2Φ(−a), a > 0
- P(a < Z < b) = Φ(b) − Φ(a) - P( | Z | < a) = 2Φ(a) − 1,a > 0
• note that:
- | Z | ≥ a ⟺ Z ≤ − a or Z ≥ a - |Z| ≤ a ⟺ − a ≤ Z ≤ a

• given α, find the critical value zα

- equivalent to be given p, but here p =1−α

• given X ∼ N(μ, σ 2), find the probability of X

X−μ
- standardize X first: Z = ∼ N(0,1)
σ

( σ σ )
a−μ b−μ
P(a ≤ X ≤ b) = P ≤Z≤
then

( σ ) ( σ )
- b−μ a−μ
=Φ −Φ

32
Exercise
• Let Z be a standard normal random variable,
- 1) calculate P( | Z − 1 | ≤ 1.5)?

- 2) If P(0 ≤ Z ≤ c) = 0.291, what is the value of c?

- 3) determine zα for α = 0.2

33
Exercise
• Let X be the IQ of a randomly selected person. Assume X ∼ N(100,256). What is
the probability that a randomly selected person has an IQ
- above 140?

- between 92 and 114?

- Find the median of X.

34
Normal Approximation to the Binomial
Distribution
• If X
∼ Bin(n, p) with np ≥ 10 and n(1 − p) ≥ 10, then X has
approximately a normal distribution N(np, np(1 − p)),

( np(1 − p) )
x + 0.5 − np
P(X ≤ x) ≈ Φ

The term 0.5 is added as a correction term for using a continuous distribution to
approximate a discrete distribution

• Comparison:
- Poisson approximation to binomial: n is large and p is small. n > 50 and np < 5
- Normal approximation to binomial: not too skewed (enough symmetry). np ≥ 10
and n(1 − p) ≥ 10.

35
Exercise
• At a particular college, the pass rate of a course is 72%. If 500
students enroll in a semester, what is the probability that at most
375 students pass?

36