Using deep learning for recommendation systems
In this section, we will learn how to use deep learning to create a recommendation system. We will use the example of a movie recommendation system to understand the underlying concepts. First, we will review what movie recommendation system data looks like, and we will then discuss a deep learning-based solution to build a recommendation system.
Understanding a movie recommendation system dataset
The dataset for building a recommendation system looks like the toy example shown in Figure 18.2. On the vertical axis, we have 5 users in our database. And on the horizontal axis, we have 8 movies.
Figure 18.2: Example of a movie database presented as a user-movie matrix, where the entries of the matrix represent the rating given by the user to the movie
For a given user and a given movie, we find a rating ranging from 1 to 5 stars, 1 being the worst and 5 being the best. This rating is used as a target to train a deep...
