Data for a typical recommender system consists of two entities in the form of “users” and “items” organized in what’s called a utility matrix. While it sounds intimidating, a utility matrix is a simple pivot table containing pairs of values, such as an individual user and their ratings for given items.


A white rectangular grid with black text  Description automatically generated

Table 2: Example of a utility matrix


Users are often arranged vertically in rows, and items are placed horizontally in columns. The preferences of each user for a given item can be found by searching for the relevant column. The aligning cell depicts the user’s preference for that item, which is generally represented ...

Get Machine Learning: Make Your Own Recommender System now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.