Given a set of transactions T , the goal of association rule mining is to find
all rules having support ≥ minsupthreshold and con f idence ≥ mincon f threshold.
The brute-force approach would be to list all possible association rules, compute
the support and confidence for each rule and then prune rules that do not satisfy
both conditions. This is, however, computationally very expensive. For this reason,
we take a two-step approach: (1) Generate all itemsets whose support ≥ minsup
(Frequent Itemset Generation); (2) Generate high confidence rules from each frequent itemset (Rule Generation)