Content-based filtering

Another popular branch of techniques is content-based filtering. The algorithms start with a description of items, and they don't need to take account of different users at the same time. For each user, the algorithms recommend items that are similar to its past purchases.

Here are the steps to perform a recommendation:

  1. Define item descriptions.
  2. Define user profiles based on purchases.
  3. Recommend to each user the items matching its profile

User profiles are based on their purchases, so the algorithms recommend items similar to past purchases.

Get Building a Recommendation System with R 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.