INBA 6152

Business Analytics and richard.coutou@fac.gsb.edu.tt

Quantitative Analysis
Models for Decision
Making sheldon.kelly@fac.gsb.edu.tt
4.
 Bayes theorem is used to incorporate additional
information and help create posterior probabilities from
original or prior probabilities
Prior Probabilities

Bayes Posterior
Process Probabilities

New Information

2014 Pearson Education 3 1-3

Probability: often used to refer to frequency
but
Bayesian Probability: a measure of a state of knowledge.
It quantifies
uncertainty. Allows us to reason using
uncertain statements.
A Bayesian model is continually updated as more data is
acquired.
 Bayes' Theorem provides a way to revise existing
predictions or theories given new or additional
evidence.
In finance, Bayes' theorem can be used to rate the
risk of lending money to potential borrowers.
 F U

H T H H

2014 Pearson Education 1-6

 Bayes' theorem gives the probability of an event
based on information that is or may be related to
that event. The formula can be used to see how the
probability of an event occurring is affected by new
information, assuming the new information is true.
 | ()
| =
()
P(A) the PRIOR PROBABILITY represents your
knowledge about A before you have gathered data.
e.g. if 0.01 of a population has schizophrenia then the
probability that a person drawn at random would
have schizophrenia is 0.01
 | ()
| =
()
P(B|A) the CONDITIONAL PROBABILITY the
probability of B, given A.
e.g. you are trying to roll a total of 8 on two dice.
What is the probability that you achieve this, given
that the first die rolled a 6?
 | ()
| =
()
So the theorem says:
The probability of A given B is equal to the probability of B
given A, times the prior probability of A, divided by the
prior probability of B.
Mode of transport: Probability he is late:
Car 50%
Bus 20%
Train 1%

Suppose that an employee is late one day.

His manager wishes to estimate the probability that he traveled to
work that day by car.

He does not know which mode of transportation the employee usually

uses, so he gives a prior probability of 1 in 3 to each of the three
possibilities.
 | ()
| =
()
P(car|late) = P(late|car) x P(car) / P(late)

P(late|car) = 0.5 (he will be late half the time he drives)

P(car) = 0.33 (this is the managers assumption)
P(late) = 0.5 x 0.33 + 0.2 x 0.33 + 0.01 x 0.33
(all the probabilities that he will be late added together)

P(car|late) = 0.5 x 0.33 / 0.5 x 0.33 + 0.2 x 0.33 + 0.01 x 0.33

= 0.165 / 0.71 x 0.33
= 0.7042
 Given that a drug test is 98% accurate - (|)
assume 0.5% of people use the drug - ()
If an employee selected at random tests positive for the drug, what is
the probability that the person is actually a user of the drug?
| ()
| =
()
| = (0.98 * 0.005) / ((0.98 * 0.005) + (0.02 * 0.995)) = 0.0049 /
(0.0049 + 0.0199) = 19.76%
This shows that even if a person tested positive in this scenario, it is
much more likely the person is not a user of the drug.
 Segmentation is the act of inferring some truth about a
business based on the information at hand.
You can segment any group of individuals customers,
distributors, vendors based on pre-existing information.
By assigning an artificial grouping to this set of individuals,
you can uncover additional information that helps
advance your understanding of how each group
contributes to your business goals.
Then, you can adjust your strategy based on that
updated understanding.
 Consider a local manufacturer executive decides to
segment the distributors of their financial products by
volume of sales, thus creating three subgroups:
1. high-volume distributors,
2. average distributors, and
3. low-volume distributors.

With this artificial grouping, analysts can then

examine the revenue and costs associated with each
group, and check against business models to learn if
a high sales volume contributes to profitability.
 For example, if analysts confirm that higher sales volume
leads to higher profit, that information can (and should)
inform the executives decision making and business
strategy. How can they best support high-volume
distributors? How can they incentivize other distributors to
sell more?
But if the analysis shows that volume is not a key indicator
of profitability (in many cases margin, rather than volume,
drives profit), then executives must update their models
based on this new information, and shift business strategy
based on an updated understanding of profitability.
 A cup contains two dice identical
in appearance but one is fair
(unbiased) and the other is
loaded (biased)
The probability of rolling a 3 on the
fair die is 1/6 or 0.166
The probability of tossing the same number on the loaded die is
0.60
We select one by chance, toss it, and get a 3
What is the probability that the die rolled was fair?
What is the probability that the loaded die was rolled?

2014 Pearson Education 1 - 17

 The probability of the die being fair or
loaded is
P(fair) = 0.50 P(loaded) = 0.50
And that
P(3 | fair) = 0.166 P(3 | loaded) = 0.60
The probabilities of P(3 and fair) and P(3 and
loaded) are
P(3 and fair) = P(3 | fair) x P(fair)
= (0.166)(0.50) = 0.083
P(3 and loaded) = P(3 | loaded) x P(loaded)
= (0.60)(0.50) = 0.300
2014 Pearson Education 1 - 18
 The probability of the die being fair or loaded is:
P(fair) The sum P(loaded)
= 0.50 of these =probabilities
0.50
gives us the unconditional
And that probability of tossing a 3
P(3 | fair) = 0.166 P(3 | loaded) = 0.60
P(3)of=P(3
The probabilities 0.083
and+fair)
0.300and= P(3
0.383
and
loaded) are:
P(3 and fair) = P(3 | fair) x P(fair)
= (0.166)(0.50) = 0.083
P(3 and loaded) = P(3 | loaded) x P(loaded)
= (0.60)(0.50) = 0.300
Copyright
2014 Pearson2015
EducationPearson 2 19 1 - 19
 If a 3 does occur, the probability that
the die rolled was the fair one is
P(fair and 3) 0.083
P(fair | 3) = = = 0.22
P(3) 0.383

The probability that the die was loaded is

P(loaded and 3) 0.300

P(loaded | 3) = = = 0.78
P(3) 0.383

2014 Pearson Education 1 - 20

 These are the revised or
posterior probabilities
If for
a 3 the
does occur,
next roll the probability
of the die that the die
rolled
We usewasthese
the fair one is our
to revise
prior probability
P(fair and 3) 0.083
estimates
P(fair | 3) = = = 0.22
P(3) 0.383

The probability that the die was loaded is

P(loaded and 3) 0.300

P(loaded | 3) = = = 0.78
P(3) 0.383

Copyright
2014 Pearson2015
EducationPearson 2 21 1 - 21
 TABLE 2.3 Event B has occurred

STATE OF P (B | STATE PRIOR JOINT POSTERIOR

NATURE OF NATURE) PROBABILITY PROBABILITY PROBABILITY

A P(B | A) x P(A) = P(B and A) P(B and A)/P(B) = P(A | B)

A P(B | A) x P(A) = P(B and A) P(B and A)/P(B) = P(A | B)
P(B)

TABLE 2.4 A 3 is rolled

STATE OF P (B | STATE PRIOR JOINT POSTERIOR

NATURE OF NATURE) PROBABILITY PROBABILITY PROBABILITY

Fair die 0.166 x 0.5 = 0.083 0.083 / 0.383 = 0.22

Loaded die 0.600 x 0.5 = 0.300 0.300 / 0.383 = 0.78
P(3) = 0.383

Copyright
2014 Pearson2015
EducationPearson 2 22 1 - 22
 We can compute revised probabilities
more directly by using
P(B | A)P(A)
P(A | B) =
P(B | A)P(A) + P(B | A)P( A)
where
A = the complement of the event A;
for example, if A is the event fair die,
then A is loaded die

2014 Pearson Education 1 - 23

 P(AB)
Conditional probabilityP(A | B) = P(B)

From the previous example
Replace A with fair die, A with loaded
die, B with 3 rolled
P(fair die | 3 rolled)
P(3 | fair)P(fair)
=
P(3 | fair)P(fair) + P(3 | loaded)P(loaded)
(0.166)(0.50) 0.083
= = = 0.22
(0.166)(0.50)+(0.60)(0.50) 0.383
Copyright
2014 Pearson2015
EducationPearson 2 24 1 - 24
 Additional information from a second
experiment
If you can afford it, perform experiments several
times
We roll the die again and again get a 3

P(fair) = 0.50 and P(loaded) = 0.50

P(3,3 | fair) = (0.166)(0.166) = 0.027
P(3,3 | loaded) = (0.6)(0.6) = 0.36

Copyright
2014 Pearson2015
EducationPearson 2 25 1 - 25
 P(3,3 and fair) = P(3,3 | fair) P(fair)
Additional information from a second
= (0.027)(0.5) = 0.013
experiment
P(3,3 and loaded) = P(3,3 | loaded) P(loaded)
If you can afford it, perform experiments several
times = (0.36)(0.5) = 0.18
We roll the die again and again get a 3

P(fair) = 0.50 and P(loaded) = 0.50

P(3,3 | fair) = (0.166)(0.166) = 0.027
P(3,3 | loaded) = (0.6)(0.6) = 0.36

Copyright
2014 Pearson2015
EducationPearson 2 26 1 - 26
 P(3,3 and fair) = P(3,3 | fair) P(fair)
Additional information from a second
= (0.027)(0.5) = 0.013
experiment
P(3,3 and loaded) = P(3,3 | loaded) P(loaded)
If you can afford it, perform experiments several
times = (0.36)(0.5) = 0.18
We roll the die again and again get a 3
P(3,3 and fair) 0.013
P(fair | 3,3) = = = 0.067
P(fair) = 0.50 and P(loaded)
P(3,3) = 0.50 0.193
P(3,3 | fair) = (0.166)(0.166) = 0.027
P(3,3 =P(3,3 and= 0.36
loaded) 0.18
P(loaded | 3,3) =
| loaded) (0.6)(0.6) = = 0.933
P(3,3) 0.193

Copyright
2014 Pearson2015
EducationPearson 2 27 1 - 27
 After the first roll of the die
probability the die is fair = 0.22
probability the die is loaded = 0.78

After the second roll of the die

probability the die is fair = 0.067
probability the die is loaded = 0.933

Copyright
2014 Pearson2015
EducationPearson 2 28 1 - 28
 It allows us to put probability values on unknowns. We can
make logical inferences even regarding uncertain
statements.

2014 Pearson Education 29 1 - 29

 A random variable assigns a real number to
every possible outcome or event in an
experiment
X = number of refrigerators sold during the day

Discrete random variables can assume only a

finite or limited set of values
Continuous random variables can assume
any one of an infinite set of values

Copyright
2014 Pearson2015
EducationPearson 2 30 1 - 30
 TABLE 2.5 Random Variable that are Numbers

RANGE OF
RANDOM
EXPERIMENT OUTCOME RANDOM
VARIABLES
VARIABLES
Stock 50 Number of Christmas
X 0, 1, 2,, 50
Christmas trees trees sold
Inspect 600 Number of acceptable
items items Y 0, 1, 2,, 600

Send out 5,000 Number of people

sales letters responding to the letters Z 0, 1, 2,, 5,000

Build an Percent of building

apartment completed after 4 R 0 R 100
building months
Test the lifetime Length of time the bulb
of a lightbulb lasts up to 80,000 S 0 S 80,000
(minutes) minutes

Copyright
2014 Pearson2015
EducationPearson 2 31 1 - 31
 TABLE 2.6 Random Variables not Numbers

RANGE OF RANDOM RANDOM

EXPERIMENT OUTCOME VARIABLES VARIABLES
Students Strongly agree (SA) 5 if SA 1, 2, 3, 4, 5
respond to a Agree (A)
4 if A
questionnaire Neutral (N) X = 3 if N
Disagree (D) 2 if D
Strongly disagree (SD) 1 if SD
One machine is
inspected
Defective
Not defective
Y= { 0 if defective
1 if not defective
0, 1

Consumers Good 3 if good 1, 2, 3

respond to how Average Z = 2 if average
they like a Poor 1 if poor
product

Copyright
2014 Pearson2015
EducationPearson 2 32 1 - 32
 End

2014 Pearson Education 33 1 - 33