Machine Learning: Make Your Own Recommender System

8

 

 

 

CONTENT-BASED FILTERING

 

As established in Chapter 2, content-based filtering methods provide recommendations based on similar item attributes and the profile of an individual user’s preferences. The content-based filtering system then attempts to recommend items similar to those that a user has liked or browsed in the past. After purchasing a book about “machine learning,” for example, Amazon’s content-based filtering is likely to serve you other books:

- from the same author,

- from the same category, e.g., data science, and

- have similar title keywords, e.g., “machine learning.”

As expected, content-based filtering relies heavily on a description of the item’s characteristics and the profiling of individual user preferences. ...