Business Intelligence Assignment Guide

1. The fitness center I go to offers a number of different classes. Since clients attend more than one type of class, the manager wishes to identify different groups of classes that clients tend to attend together. This will help her schedule classes so as to minimize overlap among related classes.
In this exercise we will concern ourselves with only seven different classes. Using a minimum support threshold of
30% and a minimum confidence level of 60%, (manually) apply association rule mining to the set of transactions given below to identify all valid rules. Clearly list out all relevant steps and report the support, confidence and lift for each valid rule that you generate.
Customer ID
1
2
3
4
5
6
7
8
9
10

Food
Yoga, Pilates,
Weight Loss, Step
Aerobics
Zumba, Cardio,
Weight Loss,
Spinning
Yoga, Zumba,
Pilates, Step
Aerobics
Yoga, Pilates, Step
Aerobics
Zumba, Cardio,
Spinning
Step Aerobics,
Spinning, Weight
Loss
Zumba, Pilates,
Yoga
Yoga, Spinning
Pilates, Step
Aerobics
Step Aerobics,
Pilates, Spinning

Solution 1]
Given:
A) Minimum Support Threshold = 30%
B) Minimum Confidence level = 60%
Applying Apriori Algorithm:



Support greater than the user-specified support threshold min_sup (minimum support) , and Confidence greater than the user-specified confidence threshold min_conf (minimum confidence) Formulae to be used:
a) Support = No of Transactions containing all items in antecedent and consequent transactions in the database.

/ No of

b) Confidence = No of Transactions containing all items in antecedent and consequent
No of transactions containing items in the antecedent.
c) Lift = Confidence of the Rule

/

/ Support of the Consequent.

1. One Element sets validity check:
Step 1 – First look for most frequent item as a single set.
1. Yoga =5/10 = 0.5 = 50% (Support)
2. Pilates = 6/10 = 0.6 = 60% (Support)
3. Weight loss = 3/10 = 0.3 = 30% (Support)
4. Step Aerobics = 6/10 = 0.6 = 60% (Support)
5. Zumba = 4/10 = 0.4 = 40% (Support)
6. Cardio = 2/10 = 0.2 = 20% (Support) [Discarded as its support is below the minimum support threshold]
7. Spinning = 5/10 = 50% (Support)
Therefore, we get only 6 items to determine the relation.
Set = {yoga, Pilates, Weight loss, Step Aerobics, Zumba, Spinning}

2. Two Element sets validity check: a] {Yoga, Pilates}

b] {Pilates, Step Aerobics} c] {Step Aerobics, Spinning} (This rule is discarded as the support of this rule is only
20% which is below the minimum support level of 30%) d] {Yoga, weight loss} (This rule is discarded as the support of this rule is only 20% which is below the minimum support level of 30%) e] {Yoga, step aerobics} f] {Zumba, spinning} (This rule is discarded as the support of this rule is only 20% which is below the minimum support level of 30%)

g] {Yoga, Zumba} (This rule is discarded as the support of this rule is only 20% which is below the minimum support level of 30%) h] {Yoga, spinning} (This rule is discarded as the support of this rule is only 10% which is below the minimum support level of 30%) i] {Pilates, weight loss} (This rule is discarded as the support of this rule is only 10% which is below the minimum support level of 30%) h] {Pilates, Zumba} (This rule is discarded as the support of this rule is only 20% which is below the minimum support level of 30%) l] {Pilates, spinning} (This rule is discarded as it has no support). m] {Weight loss, step aerobics} (This rule is discarded as the support of this rule is only
20% which is below the minimum support level of 30%) n] {Weight loss, Zumba} (This rule is discarded as the support of this rule is only 10% which is below the minimum support level of 30%) o] {Weight loss, spinning} (This rule is discarded as the support of this rule is only 20% which is below the minimum support level of 30%) p] {Step aerobics, Zumba} (This rule is discarded as the support of this rule is only 10% which is below the minimum support level of 30%) q] {Step aerobics, spinning} (This rule is discarded as the support of this rule is only 20% which is below the minimum support level of 30%)

Two Element Sets Valid Rules: To be checked
1. { YogaPilates} : Support = 40% ; Confidence = 4/5 = 80% ; Lift = 0.8/0.6 = 1.33
2. { PilatesStep aerobics} : Support = 50% ; Confidence = 5/6 = 83.33% ;
Lift = 0.833/0.6 = 1.38
3. { Yogastep aerobics} : Support = 30% ; Confidence = 3/5 = 60% ;
Lift = 0.6/0.6 = 1
4. { PilatesYoga} : Support = 40% ; Confidence = 4/6 = 66.7% ;
Lift = 0.667/0.5 = 1.334
5. { Step aerobicsPilates } : Support = 50% ; Confidence = 5/6 = 83.33% ;
Lift = 0.833/0.6 = 1.38
6. {Step aerobicsYoga} : Support = 30% ; Confidence = 3/6 = 50% ;
(Hence, not a valid rule as it doesn’t meet the required confidence.)
3. Three Item sets validity check:
1. {Yoga Pilates, Weight Loss} : Support = 10% (Doesn’t meet the minimum required support, Hence discarded)
2. { Yoga, PilatesStep aerobics} : Support = 30%, Confidence = 3/4 = 75% , Lift =
0.75/0.6 =1.25
3. {YogaPilates, Step aerobics} : Support = 30%, Confidence = 3/5 = 60% , Lift =
0.6/0.5 = 1.2
4. {PilatesYoga, Step aerobics}: Support = 30%, Confidence = 3/6 = 50% (Discarded as this doesn’t meet the minimum confidence level).
5. {Yoga, Step aerobicsPilates} : Support = 30%, Confidence = 3/3 = 100% , Lift =
1/0.6 = 1.667
6. {Pilates, Step aerobicsYoga} : Support = 30%, Confidence = 3/5 = 60% , Lift =
0.6/0.5 =1.2
7. {YogaPilates, Zumba}: Support = 20% (Doesn’t meet the minimum required support, Hence discarded).

8. {Step aerobicsYoga, Pilates}: Support=30%, Confidence= 3/6 =50% (Discarded as this doesn’t meet the minimum confidence level).
Total Valid Association rules are as listed below:
1.
2.
3.
4.
5.
6.
7.
8.
9.

{ Yoga  Pilates}
{ Pilates  Step aerobics}
{ Yoga  Step aerobics}
{ Pilates  Yoga}
{ Step aerobics  Pilates}
{ Yoga, Pilates Step aerobics}
{ Yoga  Pilates, Step aerobics}
{ Yoga, Step aerobics  Pilates}
{ Pilates, Step aerobics Yoga}

Question 2:

The following contingency table summarizes the responses obtained after surveying a group of people about their preferences for cola and cheese.
Cheese
Cola
No Cola

(a)

No Cheese
1500
800
2300
1800
3800
2600

2300
4100
6400

no cola" is mined. Given a minimum support threshold of 35% and a minimum confidence threshold of 40%, is this association rule valid?

(b) Based on the given data, correct? Clearly explain how steps. (1 Point)
a. Preferences for cheese
b. Preferences for cheese negatively correlated
c. Preferences for cheese positively correlated

which one of the following statements is you arrived at the result detailing all relevant and no cola are statistically independent and no cola are statistically dependent and and no cola are statistically dependent and

Solution 2]
A)
Cola
No Cola
Sum

Cheese
1500
2300
3800

No Cheese
800
1800
2600

sum
2300
4100
6400

I.

II.

Considering minimum support threshold of 35% and a minimum confidence threshold of
40% for "cheese => no cola" .
Support = 2300/ 6400 = 0.359375 = 35.9375 %
Hence it is meeting the minimum threshold level.
Confidence:
2300/3800= 0.605263 = 60.5263%.
Hence it is meeting the minimum confidence level.
Hence the association rule "cheese  no cola" is Valid.

B)
Considering Statistical (In) Dependence formulae
S=Cheese and B= No Cola
P (S and B) = 2300/6400 = 0.3593
P (S) = 3800/6400 = 0.59375
P (B) = 4100/6400 = 0.640625
P(S) * P (B) = 0.380
We see that P (S and B) < P(S) * P (B) and hence our answer choice would be OPTION
B i.e.
Preferences for cheese and no cola are statistically dependent and negatively correlated