In the Association Rules model, the minimum confidence percent should be 90 (or 0.9).

• A. True
• B. False
• C. Not specified
• D. Depends on dataset

Answer: (B)

Confidence in an association rule is the probability that the rule’s consequent occurs given its antecedent, calculated as support(A ∪ B) divided by support(B). The minimum confidence threshold is a user-chosen setting, not a universal rule. Requiring a 90% confidence would be too strict for many datasets, causing many potentially useful rules to be missed. The appropriate threshold depends on the data, its size and sparsity, and the trade-off you want between precision and coverage. In practice, analysts often start with lower confidence and then prune rules using additional measures like lift or by considering domain knowledge. So there isn’t a universal requirement that the minimum confidence must be 90%.