Association Rules

Identify relationships, patterns or associations between variables in large datasets

$Y\to Z$

• $Y$ is the Antecedent or LHS
• $Z$ is the Consequent or RHS
• $Y$ and $Z$ must come from the same collection of items
• $Y$ and $Z$ must not overlap

If $Y$ is in a transaction then $Z$ is also likely to be present in the same transaction

$I$ is the Universal Set or the collection of all possible items.

An itemset is a collection of one or more items from the universal set.

An itemset containing $k$ items is called a $k$-itemset