Association Rules

Association Rules are frequent if-then associations that indicate probability of relationships between data items and are used to find correlations and co-occurrences between them.
Association Rules are either created by humans or is generated using automatic Rule Induction via Rule-Based Machine Learning (RBML) or similar methods, and they are considered Unsupervised Learning algorithms.

Each Rule often has two data items as it's components:

  • Antecedent (if)
  • Consequent (then)

Metrics:

  • Support Criteria: Indicates how frequently the items appear in the data.
  • Confidence Criteria: Indicates the number of times the if-then statements are found true.
  • Lift Criteria: It's the ratio of confidence to support, indicating how many times an if-then statement is expected to be found true.

Algorithms: