Although the inner workings of recommendation systems may seem intimidating at first, they are actually quite intuitive. Let's take an example of various movies and users. Each user has the option to rate a movie on a scale of 1 to 5. The recommendation system will try to find users with similar preferences to a new user, and will then recommend movies that the new user will probably like, as similar users also like them. Let's take the following simple example, consisting of four users and six movies:
|
User |
Interstellar |
2001: A Space Odyssey |
The Matrix |
Full Metal Jacket |
Jarhead |
Top Gun |
|
U0 |
5 |
4 |
2 |
1 |
||
|
U1 |
1 |
4 |
4 |
3 |
||
|
U2 |
4 |
4 |
1 |
|||
|
U3 |
4 |
5 |
5 |
4 |
Ratings for each movie from each user
As is evident, each user has rated a number of movies, although not all users watched the same...
