B04 Association Rules

Question 1

__________ are involved with summarizing or grouping data in new and interesting ways. In these types of models, no single feature is more important than the other. The process of training a __________ is known as ___________

Answer 1

Descriptive Models

Descriptive Model

Unsupervised Learning

Question 2

The application of machine learning to customer retail transactional data in order to detect patterns in purchasing behavior.

Answer 2

Market Basket Analysis

Question 3

The result of a market basket analysis is a collection of __________ that specify patterns found in the relationship among items or itemsets.”

Answer 3

Association Rules

Question 4

{beer, milk}→{diaper} LHS -> RHS Define the LHS and RHS

Answer 4

LHS condition that needs to be met to trigger the rule RHS the expected result of meeting the condition

Question 5

Association Rules imply causality between items True False

Answer 5

False

Question 6

Kinds of Association Rules 1.______ 2.______ 3.______

Answer 6

Actionable Trivial Inexplicable

Question 7

Define Itemset

Answer 7

A collection of one or more items e.g. {beer, bread, diaper, milk, eggs}.

Question 8

Define Transaction

Answer 8

The itemset for an observation. e.g. T(3) = {milk, diaper, beer, coke}

Question 9

Define Support Count (σ)

Answer 9

Frequency of an itemset. e.g. σ({beer, milk, diaper}) = 2

Question 10

Define Support (s)

Answer 10

How frequently a rule occurs in the dataset.

Support is the fraction of transactions containing an itemset.

Question 11

Define Confidence

Answer 11

The predictive power or accuracy of a rule.

Confidence is the support of the itemset containing both X and Y divided by the support of the itemset containing only X.

Question 12

Define Lift

Answer 12

The increased likelihood that a rule occurs in a dataset relative to its typical rate of occurrence.

Lift is the confidence of the itemset containing both X and Y divided by the support of the itemset containing only Y.

Question 13

Rules with ___ support may occur simply by chance and are typically not actionable

Answer 13

Low

Question 14

The Apriori Principle is based on the __________ property of support.

Answer 14

anti-monotone

Question 15

The support of an itemset _____ (always, sometimes, never) exceeds that of its subsets.

Answer 15

Never

Question 16

How do you reduce the computational complexity of Association rules?

Answer 16

1. Apriori Algorithm - pruning the itemset lattice
2. FP-growth - Reduce number of comparisons by using advanced data structures to store the candidate itemsets or to compress the dataset

Question 17

Approaches to Association Rules include:

Answer 17

1. Brute Force Approach - generate every itemset (computationally expensive)
2. Frequent Itemset Generation - generate every itemset whose support is greater than an established minimum (computationally expensive)
3. Apriori approach
4. FP Growth

Question 18

How does the Apriori approach work?

Answer 18

Pruning - If we identify an itemset as being infrequent, then its supersets should not be generated/tested

Question 19

Advantages of Association Rules?

Answer 19

1. Capable of working with large amounts of transactional data. 2.Rules are easy to understand.
2. Useful for “data mining” and discovering unexpected patterns in data.

Question 20

Weaknesses of Association Rules?

Answer 20

• Not very useful for small data sets.
• Separating true insight from common sense requires some effort.
• Easy to draw misleading conclusions from random patterns.