Association Mining for Machine Learning

Discover three storage strategies for transaction databases in association mining—simple storage, horizontal storage, and vertical storage—and learn to compute item frequencies and intersections.

Learn the naive algorithm for finding association rules via brute-force item pair and triple evaluation, using a 4-transaction dataset with 50 percent support and 75 percent confidence.

Learn about the apriori algorithm for finding association rules, its two-phase process of identifying frequent item sets and generating rules, and its role in modern recommendation systems.

Explore the apriori algorithm's two-phase approach to association mining. Identify frequent one-item sets and generate candidate two-item sets, pruning to pairs such as bread with jam and bread with cornflakes.

Apply the Apriori algorithm’s second phase to generate association rules from two-item pairs. Compute the confidence values and confirm four rules meet the threshold with 75% to 100% confidence.

The lecture explains the apriori algorithm for mining association rules using a four-transaction dataset, showing phase one frequent itemsets and phase two rule generation with 50% support and 70% confidence.