Introduction to Market Basket Analysis with the Apriori Algorithm
In the previous example, we analyzed the ratings provided by different customers in order to perform mixed segmentation. However, sometimes, a company only has complete knowledge about the set of products bought by its customers. More formally, given a set P = {p1, p2, …, pn} of products, a transaction, Ti, is a subset of P:
A collection of transactions (often called a database) is a set of subsets, Ti:
The main goal of market basket analysis is to mine all existing association rules that can be expressed in the generic form:
To avoid confusion, the previous expression means that, given a transaction containing a set of items, the probability of finding the item pt is greater than a discriminant threshold, (for example, 0.75). The value of this process is straightforward because a company can optimize its offers based on the evidence provided by the actual transactions. For example, a retailer...