Association Rules and Apriori

Question 1

What are frequent patterns in data commonly referred to as?

Answer 1

Frequent patterns are also known as association rules.

Question 2

What are the two types of analysis done using frequent patterns?

Answer 2

1. Frequent itemsets which leads to the discovery of associations and correlations among items. 2. Frequent subsequences which allows the discovery of patterns across time or positions in a dataset.

Question 3

What is Market Basket Analysis?

Answer 3

Market Basket Analysis is the type of analysis that identifies sets of items that appear together in transactional datasets such as which wines are sold together with which dish in a restaurant.

Question 4

What is the Apriori Algorithm used for?

Answer 4

The Apriori Algorithm is used to identify frequent itemsets in a dataset.

Question 5

Name two other frequently used algorithms aside from Apriori.

Answer 5

1. FPGrowth 2. Eclat.

Question 6

What is the objective of frequent substructures?

Answer 6

The objective of frequent substructures is to find interesting subgraphs in data which can be combinations of frequent itemsets and frequent subsequences.

Question 7

Define frequent subsequences in the context of datasets.

Answer 7

Frequent subsequences refer to the discovery of patterns across time or positions in a dataset such as the sequential order of purchasing history.

Question 8

What are some algorithms used for exploring sequences?

Answer 8

1. GSP 2. Spade 3. PrefixScan.

Question 9

What issues can arise with association rules?

Answer 9

Problems include redundancy too many rules making it difficult to find interesting patterns and too few rules if minimum support or confidence thresholds are too high.

Question 10

What are actionable rules in association rules context?

Answer 10

Actionable rules contain high-quality actionable information that can lead to insights and actions.

Question 11

What is the definition of support in association rules?

Answer 11

Support is the fraction of rules that occur in all observations.

Question 12

What does confidence measure in the context of association rules?

Answer 12

Confidence is the probability of a rule being correct for a new observation.

Question 13

How is lift defined in association rules?

Answer 13

Lift is the ratio by which the confidence of a rule exceeds the expected confidence indicating how much more likely the right-hand side occurs when the left-hand side is true.

Question 14

What is required for data to find frequent subsequences?

Answer 14

1. A timestamps or sequencing information to determine when transactions occurred relative to each other. 2. Identifying information such as customer ID to know which transactions belong to the same entity.

Question 15

How are rules generated from frequent itemsets?

Answer 15

Rules are generated by calculating confidence and removing itemsets that do not meet the parameter criteria.

Question 16

What is the significance of pruning in the Apriori Algorithm?

Answer 16

Pruning involves removing items or itemsets that do not reach the minimum support which helps streamline the analysis to only relevant data.

Question 17

What is the typical structure of association rules?

Answer 17

Association rules consist of two parts: Antecedents (Left-Hand-Sides LHS) and Consequents (Right-Hand-Sides RHS).

Question 18

What are common applications of association rules in business?

Answer 18

Common applications include cross-selling upselling product placement and understanding customer behavior.

Question 19

What is the disadvantage of the Apriori Algorithm?

Answer 19

Frequent itemsets require the entire dataset to be scanned which can be computationally expensive if itemsets are very large and minimum support is very low.

Question 20

Explain the concept of redundant rules in association analysis.

Answer 20

Redundant rules are rules that include other rules which do not provide new insights and should be removed to reduce complexity.

Question 21

What are typical parameters set in association rule mining?

Answer 21

Typical parameters include Minimum Support Minimum Confidence and Maximum Itemsets Length.

Question 22

What does ‘frequent subgraphs’ refer to?

Answer 22

Frequent subgraphs refer to common patterns in data that are represented as graphs which help to identify relationships between items.

Question 23

How does sequential pattern analysis differ from traditional association rule analysis?

Answer 23

Sequential Pattern Analysis considers not only associations but also the sequences of items where the order in time is important.

Question 24

What type of rules are classified as trivial rules?

Answer 24

Trivial rules are those that are already known by anyone familiar with the business such as well-established associations.

Question 25

What does the term ‘inexplicable rules’ refer to in association analysis?

Answer 25

Inexplicable rules are those that seem to have no explanation and do not imply any clear course of action.

Question 26

Describe the final step in the Apriori Algorithm process.

Answer 26

The final step is to generate rules by calculating confidence based on frequent itemsets and pruned lists retaining only those that meet the established parameters.