0% found this document useful (0 votes)
43 views33 pages

Lecture6 - Random Variable - 0925

Uploaded by

九.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
43 views33 pages

Lecture6 - Random Variable - 0925

Uploaded by

九.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 33

DOTE 2011 | Fall 2024

@ CUHK Business School

Statistical Analysis for Business Decisions


Random Variable

Yunduan Lin
Assistant Professor
Department of Decisions, Operations and Technology
CUHK Business School
Agenda

Statistical Analysis for Business Decisions

01 Continous random variable


o Uniform distribution
o Exponential distribution
o Normal distribution

02 Distribution in Excel
o Function
o Calculation
o Plot
Continuous Random Variable - Definition

Continuous Random Variable

o the support is infinite (uncountable)


o naturally arises when the number of outcomes in the sample space is infinite (e.g., temperature)
o usually assign real numbers to each outcome (e.g., temperature, stock price)

the set of nonnegative real numbers


Example: Stock market index the set of real numbers

o Sample space
o Let X be a random variable of the stock market index
o Support of X is the same as S
Uniform Distribution
Recall:

Uniform distribution with two outcomes


x 0 1
P(x) 1/2 1/2
F(x) 1/2 1

Uniform distribution with three outcomes

x 0 1 2
P(x) 1/3 1/3 1/3
F(x) 1/3 2/3 1
Uniform Distribution
Move from discrete to continuous:
Now consider a uniform distribution of many outcomes.

o We can still obtain F(x) by stacking P(x).


o What happens when we have infinitely many outcomes---we have a continuous random variable.
o F(x) is still stacking over P(x) until x but now it is just adding up all the area up to x. This is
exactly integration!
o The same logic applies but we call it probability density function f(x) instead of probability mass
function P(x) because the probability at one point is so thin (1/n is so small) that it is now called
density (instead of mass)!
Cumulative Distribution Function - Definition

Cumulative Distribution Function (CDF)

Useful property:

o Denote a and b some constants such that


Probability Density Function - Definition
Although we cannot say probability at one value using P(x), we have something similar for a
small interval:

Probability Distribution (or Density) Function (PDF)

If b-a>0 is small, and

Alternatively, for some very small Δ,


PDF and CDF

Area under PDF f(x) is probability

Integration

o When we have continuous random


variable, the probability of a random
variable exactly equal to a particular value
is very small.
o When we are asking the probability that a
random variable falls within a specific
range, we are combining probabilities over
all the points in the range. Since there are
many points in a range, we need to use the
tools of integration.
PDF and CDF
For a continuous random variable:

o Pr(X=a) = Pr(X=b) = 0 for any particular values of a and b.

o
Expectation and Variance

Expectation

Variance

Note that they are the same as the discrete counterparts after we replace the sum by the integral
and add “dx”.
Uniform Distribution - Definition

Uniform distribution
Equally likely to be any value between a and b

o Support of X: [a,b] where


o Probability distribution function

o Cumulative distribution function


Uniform Distribution - Property

Uniform distribution

o Mean

o Variance

Intuition:

o Mean: the middle of the interval.


o Variance: proportional to the squared range
Uniforms Distribution - Property

o Expectation

o Variance
Uniform Distribution - Example
Example:

The daily rainfall follows a uniform distribution between 1mm and 4mm.

o What is the probability that the rainfall is between 1mm and 3mm?

Note that we have a=1, b=4

o What is the expected daily rainfall and its variance?


Exponential Distribution - Definition

Exponential distribution

o Waiting time until a success in Poisson with mean


o Memoryless: same probability after waiting
o Support of X:
o Probability distribution function

o Cumulative distribution function


Exponential Distribution - Property

Exponential distribution In some definitions, people will define

o Mean

o Variance

Intuition:
When we expect λ of events happen in an hour,
then we expect 1/λ is the arrival of the first event
Exponential Distribution - Example
Example:

Suppose the time for next customer to arrive follows exponential distribution with mean 20 minutes.

o What is the probability that a customer arrive in within next 10 minutes?

Note that mean=20 implies that or

o What is the probability that a customer arrive in between next 5 minutes and 15 minutes?
Exponential Distribution - Example
Example:

Suppose a customer service representative in a call center takes an average of 5 minutes to handle
a customer call, and this duration follows exponential distribution.

o What is the probability that a customer call is completed within 10 minutes?

Note that mean=5 implies that or


Exponential Distribution - Example
Example:

In manufacturing or maintenance operations, the exponential distribution can be used to model the
time between equipment failures. Average time between failures: 100 hours

o What is the probability of the machine functioning without failure for a specific period (e.g., 200
hours):

Note that mean=100 implies that or


Normal Distribution - Definition

Normal distribution

o Bell curve
o Support of X:
o Probability distribution function

o Cumulative distribution function: no functional form---required calculator or table to find F(x)


Normal Distribution - Property

Normal distribution
Change the location of distribution
o Mean
Change the spread of distribution
o Variance

o Shorthand notation

Intuition:
Skewness is always zero
Kurtosis is always 3 -> excesskurtosis is always 0
Normal Distribution - Property

Linear transformation

If , then

Intuition:
Still normal distribution after a linear transformation
o Adding a constant changes only the mean: parallel shifting
o Multiplying a constant is changing only mean and variance: shifting and scaling

Standardization

If , then

Z is called “standard normal”: how many standard deviations is X from mean?


Normal Distribution - Standardization
Why standardization:

Since there is no closed form solution for F(x), we need to use numerical methods.
o In the days without computer, it would be a lot of computations. It would be easier if
calculations can be done in advanced so that we can check tables.
o Yet, as normal distribution has two parameters, we need one table for each
combination of .
o With standardization, we only need one table.
o For , if we want to calculate Pr(X<a), we only need to consider this
standardized random variable , because
Normal Distribution - Standardization

Z-table

Z-table gives F(z) = Pr(Z < z) with two versions:


o Version 1: z ≤ 0.
o Version 2: z ≥ 0.
Since f(Z=z) is symmetric about 0, two versions are interchangeable: Pr(Z < −z) = 1−Pr(Z<z)

Pr(Z<1.4) = 0.9192

Pr(Z<-1.6) = 1-Pr(Z<1.6) = 1-0.9452 = 0.0548

Pr(-1.6<Z<1.4) = Pr(Z<1.4)-Pr(Z<-1.6) = 0.9192-0.0548 = 0.8644


Normal Distribution - Standardization
The z-table can be found here

Rule of thumb
/ Three sigma rule
/ 68-95-99 rule
Normal Distribution - Inverse Normal
Use the Z-table inversely:

o Find z such that Pr(Z<z) for any given


o By complement rule, we can find z such that Pr(Z>z) for any given
o Denote such that for the given

Inverse function Inverse function

Recall that , we can write

It is the same as
Standardization - Example
Example:

Let Z follow a standard normal distribution.


o What is z such that Pr(Z<z)=0.95?

z=1.65

o What is z such that Pr(Z>z)=0.025?

F(z)=1-0.025=0.975
z=1.96
Standardization - Example
Example:

Let X follow normal distribution with

What is the probability that X is less than 16?


Standardization - Example
Example:

Let X follow normal distribution with

What is the probability that X is greater than 24?


Standardization - Example
Example:

Let X follow normal distribution with

What is the probability that X between than 16 and 24?


Sum of Normal Distribution

Sum of independent normal is also normal

If , and they are independent, then

If , and they are independent, then


Excel

Generate the Probability of Specific Distribution

See 3-Random Variables.xlsx for the following content


A Feedback Form for the Entire Term

https://docs.google.com/forms/d/e/1FAIpQLSfsEgnMFLypI_KW6GF7j_FXtVY5E4Jrmf2P_BDwaG8GXWDc0A/viewform?usp=sf_link

You might also like