Mathematics Mod.

II
2.1 Probability Calculus
30063 BIEM Class 16
Lesson 5 D Fein
A.Y. 2023-2024
• Continuous random
variables
• Notable continuous
distributions
• Continuous uniform
• Exponential
• Normal https://forms.gle/8j4ePgCHgCVx6nU19
Continuous random variables: everywhere!
There are many situations that can be represented using continuous
random variables, including
• Time spent waiting in line at the post office
• Height of a football player
• Time between customer arrivals
• Can you think of others?
Continuous random variables
Definition A random variable 𝑋: Ω → ℝ is continuous if there exists a
non-negative function 𝑓: ℝ → ℝ such that the cdf of 𝑋 can be written
as:
𝑥
𝐹 𝑥 = න 𝑓 𝑡 𝑑𝑡
−∞

Then the function 𝑓 is called the probability density function (pdf) of 𝑋.

Continuous random variables: pdf and cdf
The pdf 𝑓(𝑥) The cdf 𝐹(𝑥)
• The integral function of the pdf.
𝑓 𝑥 ≥ 0 ∀𝑥 ∈ ℝ • If the pdf 𝑓 is continuous, then
+∞ the cdf 𝐹 is derivable.
න 𝑓 𝑥 𝑑𝑥 = 1 • For all points at which the pdf 𝑓
−∞
is derivable, 𝐹 ′ 𝑥 = 𝑓 𝑥
Interpreting the pdf, 𝑓(𝑥)
Consider the probability that 𝑋 takes on in the interval (𝑥, 𝑥 + ℎ]
where the pdf 𝑓 𝑥 is a continuous function, 𝑃 𝑥 < 𝑋 ≤ 𝑥 + ℎ , ℎ > 0
Example
Find the cdf 𝐹 𝑥 for the following continuous (uniform) random
variable with pdf
1
𝑓 𝑥 = ቐ4 𝑥 ∈ 2,6
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Example – continued
Graphs of the pdf and cdf, respectively.
Continuous uniform distribution
Suppose we know nothing about a continuous random variable except
the range in which it may vary.
• We now find the density function that satisfies the normalization
condition.
• Given that we know 𝑋 can vary in a range 𝑎, 𝑏 ⊆ ℝ
𝑘 𝑥 ∈ 𝑎, 𝑏
𝑓 𝑥 =ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Compute 𝑘 such that the normalization condition is satisfied.
Continuous uniform distribution: pdf
Definition A random variable 𝑋 has a continuous uniform distribution
with parameters 𝑎, 𝑏 ∈ ℝ where 𝑎 < 𝑏, if its probability density
function is:
1
𝑓 𝑥 = ቐ𝑏 − 𝑎 𝑥 ∈ 𝑎, 𝑏
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Notation: 𝑋~𝑈 𝑎, 𝑏
Example
Let 𝑋~𝑈(2,8).
Write the density function 𝑓(𝑥) and compute the following:
1. 𝑃 𝑋 < 6
2. 𝑃 𝑋 ≥ 6
3. 𝑃(−2 < 𝑋 < 5)
𝑋~𝑈(𝑎, 𝑏): cdf
Starting with the pdf 𝑓(𝑥), compute the cdf 𝐹 𝑥
1
𝑓 𝑥 = ቐ𝑏 − 𝑎 𝑥 ∈ 𝑎, 𝑏
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑋~𝑈 𝑎, 𝑏 : graphs
1 𝑥
0 𝑥<𝑎
𝑥 ∈ 𝑎, 𝑏 𝑥−𝑎
𝑓 𝑥 = ቐ𝑏 − 𝑎 𝐹 𝑥 = න 𝑓 𝑡 𝑑𝑡 = ൞ 𝑥 ∈ 𝑎, 𝑏
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 −∞ 𝑏−𝑎
1 𝑥>𝑏
Example: piecewise-defined
Given that 𝑋 has constant densities on each of the intervals shown,
write and graph 𝑓 𝑥 .
𝑃 0 < 𝑋 ≤ 2 = 0.4
𝑃 2 < 𝑋 ≤ 4 = 0.1
𝑃 4 < 𝑋 ≤ 9 = 0.5
*Conditional probability
Before introducing the next distribution, we recall that the conditional
probability of 𝐴 given 𝐵 is defined as follows:
∀𝐴, 𝐵 ∈ Ω such that 𝑃 𝐵 ≠ 0,
𝑃 𝐴∩𝐵
𝑃 𝐴𝐵 =
𝑃 𝐵
Example
Roll two fair, 6-sided dice and take the sum the faces.
1. What is the probability that an 8 is rolled?
2. Given that the first die shows a 5, what is the probability that an 8 was
rolled?
Die 1\Die 2 𝟏 𝟐 𝟑 𝟒 𝟓 𝟔
𝟏 2 3 4 5 6 7
𝟐 3 4 5 6 7 8
𝟑 4 5 6 7 8 9
𝟒 5 6 7 8 9 10
𝟓 6 7 8 9 10 11
𝟔 7 8 9 10 11 12
Tired of waiting in line?
Let 𝑋~𝑃𝑜(𝜆) where 𝜆 is the mean number of people arriving per unit
time.
A decision maker (DM) wants to model the waiting time until the first
arrival. i.e. 𝑃 𝑇 ≤ 𝑡 = 𝐹 𝑡 .
Write the distribution for 𝑇.
Still tired of waiting?
Exponential distribution
Definition A random variable 𝑋 has an exponential distribution with
parameter α > 0, if its probability density function is:

𝛼𝑒 −𝛼𝑥 𝑥≥0
𝑓 𝑥 =ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Notation: 𝑋~𝐸𝑥𝑝 𝛼

𝑥 0 𝑥<0
The cdf is 𝐹 𝑥 = ‫׬‬−∞ 𝑓 𝑡 𝑑𝑡 = ቊ
1 − 𝑒 −𝛼𝑥 𝑥≥0
𝑋~𝐸𝑥𝑝 𝛼 : graphs
𝛼𝑒 −𝛼𝑥 𝑥≥0 0 𝑥<0
𝑓 𝑥 =ቊ 𝐹 𝑥 =ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 1 − 𝑒 −𝛼𝑥 𝑥≥0
Exponential distribution: normalization
Show that the normalization condition holds for 𝑋~𝐸𝑥𝑝 𝛼
Example
Suppose the length of time to be helped by a B-in-Touch representative
1
is an exponential random variable with parameter 𝛼 = .
30
(We will learn shortly that this corresponds to an average wait time of ...)
Suppose that a fellow student just submitted a ticket.
1
𝑋~𝐸𝑥𝑝
30
1. What is the probability you will wait less than 10 minutes?
2. What is the probability you will wait more than 1 hour?
Forgetful
Given 𝑋~𝐸𝑥𝑝(𝛼), calculate the following:

𝑃 𝑋 >𝑎+𝑏 𝑋 >𝑎 =
Memoryless
Property (no memory) For 𝑋~𝐸𝑥𝑝(𝛼), the following holds:

𝑃 𝑋 >𝑎+𝑏 𝑋 >𝑎 =𝑃 𝑋 >𝑏

Interpretation
• Let 𝑋 be the time (hours) it takes a piece of equipment to fail after it
has been installed. Suppose further that 𝑋~𝐸𝑥𝑝 𝛼 .
• Then this equation states that given the equipment has NOT failed by
time 'a,' that the probability it will fail in the next 'b' hours is exactly
the same as in the first 'b' hours.
• This implies that the aging of the equipment does not affect its
probability of failure in a given length of time.
• Thus, the exponential random variable is a 'forgetful' or 'memoryless'
distribution.
No memory property: applications
In reliability theory, random variables are used to model lifespans.

An important property of a device/component is its failure rate.

There are 3 general possibilities:

1. Increasing failure rate
2. Decreasing failure rate
3. Constant failure rate

Can you think of an example for each?

Example
An Apple Watch is guaranteed to last 5 years from the date of
manufacture. Suppose that the longevity 𝑋 of the watch is distributed
exponentially with 𝛼 = 0.5.

Your watch was manufactured at least 1 year ago. What is the

probability it will last another 3 years?
The normal distribution
The normal distribution, also called the Gaussian distribution, is a very
important distribution that arises naturally in many areas.

Gauss used it to describe the distribution of errors made while taking

repeated measurements.

Many real-world phenomena are distributed (approximately) normally:

• Biological variability among a uniform population
• Distribution of the velocities of gas molecules (Maxwell's Law)
• Etc.
The normal distribution
Definition A random variable 𝑋 has a normal distribution with parameters 𝜇
and 𝜎 2 , if its probability density function is:

1 (𝑥−𝜇)2
−
𝑓 𝑥 = 𝑒 2𝜎2
𝜎 2𝜋
where 𝑥, 𝜇, 𝜎 ∈ ℝ and 𝜎 > 0
Notation: 𝑋~𝑁 𝜇, 𝜎 2
The cumulative distribution function is:
𝑥
1 (𝑡−𝜇)2
−
𝐹 𝑥 =න 𝑒 2𝜎2 𝑑𝑡
−∞ 𝜎 2𝜋
The normal distribution: graph
𝑋~𝑁 0,1 𝑋~𝑁 3,0.4
The normal distribution

For 𝑋~𝑁(0,1), the standard normal For 𝑋~𝑁(𝜇, 𝜎 2 ), a generic normal

distribution: distribution:
• 𝑃 −1 < 𝑋 < 1 = 0.683 • 𝑃 𝜇 − 𝜎 < 𝑋 < 𝜇 + 𝜎 = 0.683
• 𝑃 −2 < 𝑋 < 2 = 0.954 • 𝑃 𝜇 − 2𝜎 < 𝑋 < 𝜇 + 2𝜎 = 0.954
• 𝑃 −3 < 𝑋 < 3 = 0.997 • 𝑃 𝜇 − 3𝜎 < 𝑋 < 𝜇 + 3𝜎 = 0.997
𝑋~𝑁(0,1) pdf and cdf graph
𝑥2 𝑥 𝑡2
1 − 1 −
𝑓 𝑥 = 𝑒 2 Φ 𝑥 =න 𝑒 2 𝑑𝑡
2𝜋 −∞ 2𝜋
 erf(𝑥)
The error function is defined as:

2 𝑥 −𝑡 2
erf 𝑥 = න 𝑒 𝑑𝑡 , −∞ < 𝑥 < +∞
𝜋 0

And the standard normal distribution function is often written as:

1 𝑥
Φ 𝑥 = 1 + erf , −∞ < 𝑥 < +∞
2 2
BIEM 16’s heights☺ (z-scores)