Software that makes viewing and product recommendations by analyzing the music, movie, merchandise and news preferences of a large group of users. For example, if many people like both A and B, any user interested in one should like the other.
Collaborative or Content-Based Filtering
To create a match, recommendation algorithms use collaborative and content-based filtering. Collaborative filtering uses direct user feedback such as ratings, reviews, the type and duration of content they watch, as well as products they purchased.
Content-based filtering takes into account the users themselves. Essential when new users come on board with no behavior history, the person's age, gender and country are used to match to a product or service. In practice, recommendation systems employ both methods. See
social shopping.