Collaborative filtering
Collaborative filtering is a famous algorithm that is based on the likings or the behavior of other users or peers unlike the content-based filtering that we studied in the previous section.
Collaborative filtering:
If the user likes some of the things that other users or peers have shown an inclination to, then the preferences of these users can be recommended to the desired user
It is referred to as the "nearest neighbor recommendation"
To implement collaborative filtering, some assumptions are made:
Likings or the behavior of peers or other users can be taken into consideration to understand and predict for the desired user. Therefore, an assumption is made that the desired user has similar tastes as the other users taken into consideration here.
If the user got a recommendation in the past based on ratings of a group of users, then the user would have a similar taste with that group.
There are different types of collaborative filtering:
Memory-based collaborative filtering...