Major Pattern Discovery Tasks: Association and Sequential Patterns

Pattern mining searches data for recurring relationships. Association analysis identifies frequent itemsets and rules; sequential pattern mining considers order and time.

Association measures

• Support measures how often an itemset occurs.
• Confidence measures how often a rule's consequent occurs when its antecedent occurs.
• Confidence-based pruning can eliminate extensions that cannot meet a required threshold.

Sequential constraints

Sequential discovery may restrict a pattern using a window size or maxgap, limiting the allowable distance between events. Always state the support/confidence thresholds and sequence constraints so the discovered pattern is reproducible and meaningful.