Which data mining method does Amazon.com use to recommend additional products when you search or buy something on its website?

• A. Decision Trees
• B. Association Rules
• C. Clustering
• D. Regression

Answer: (B)

Association rules capture how items tend to appear together in customer transactions, which is perfect for recommending additional products as you search or buy. The idea is to mine frequent patterns from baskets of items and turn them into if-then rules like: if a customer buys X, they are likely to buy Y. These rules are evaluated with metrics such as support (how often the combination occurs) and confidence (how often Y appears when X appears), and sometimes lift (how much more likely Y is with X than by chance). This approach directly models co-occurrence relationships between products, enabling real-time suggestions like “customers who bought this also bought that.”

Decision trees predict a target variable from features, not the co-occurrence relationships between items. Clustering groups similar items or users but doesn’t produce the actionable cross-item rules used for recommendations. Regression predicts a numeric value and doesn’t specify dependencies between different products.