Collaborative filtering
Collaborative filtering algorithms do not need detailed information about the user or the items. They build models based on user interactions with items such as song listened, item viewed, link clicked, item purchased or video watched. The information generated from the user-item interactions is classified into two categories: implicit feedback and explicit feedback:
- Explicit feedback information is when the user explicitly assigns a score, such as a rating from 1 to 5 to an item.
- Implicit feedback information is collected with different kinds of interaction between users and items, for example, view, click, purchase interactions in the Retailrocket dataset that we will use in our example.
Further collaborative filtering algorithms can be either user-based or item-based. In user-based algorithms, interactions between users are focused on to identify similar users. Then the user is recommended items that other similar users have bought or viewed. In item-based algorithms...
