We have all read the sentence in almost every online store: People who bought this product also bought.... It all started with Amazon in the 1990s, and it is widespread today. This same principle is even being tested in physical stores, where customers can get personalized suggestions on which items to buy based on their shopping history and similarity with other products. These techniques are generally known as cross-selling, and they are useful since it is often easier to sell an additional product to an existing customer than to acquire a new one.
In this section, we will show a simplified example of Market Basket Analysis, which is the type of data mining technique behind these recommender systems. The results might not be as accurate as those obtained with more advanced methods, but are still useful to explain the method...