Association Rules Analysis
Market basket metrics (support, confidence, lift) showing how strongly buying one product implies buying another.
Filter Rules
Top Rules by Lift (strongest product affinities)
Association Rules
| Antecedent (If customer buys) | Consequent (They likely also buy) | Support | Confidence | Lift | Conviction | Co-occurrence |
|---|---|---|---|---|---|---|
| Loading... | ||||||
Understanding the Metrics
Support = fraction of all receipts that contain both A and B. Higher = more common combination.
Confidence = P(B given A) = among receipts with A, what fraction also have B. Higher = stronger implication.
Lift = confidence / P(B). Above 1.0 means A and B are positively correlated. Below 1.0 means negative. At 1.0 they're independent.
Conviction = how much the rule fails without the association. Higher = stronger rule.