Let \(\) be a set of items, where each item is a literal. A transaction \(T \subseteq \) is a subset of \(\) . Association rules are defined on a set of transactions \(\) .
An association rule R is in the form of X → Y , where X and Y are two sets of items, that is, \(X,Y \subseteq \) . R is associated with two measures, the support sup(R) and the confidence conf (R). The support sup(R) is the probability that X appears in a transaction in \(\) . The confidence conf (R) is the conditional probability that when X appears in a transaction, Y also appears.
The concept of association rules were firstly proposed by Agrawal et al. [1] for market basket analysis. A well known illustrative example of association rules is “Diaper → Beer” which can be explained by the fact that, when dads buy diapers for their babies, they also buy beer at the same time for their weekends game watching.
Apriori, an efficient algorithm.
