In the previous recipe, we learned about generating predictions for movies that a user is likely to watch. One of the limitations of the previous way of generating predictions is that the variety of movie recommendations would be limited if we did not perform further processing on top of the movie predictions.
A variety of recommendations is important; if there were no variety, only certain types of products would be discovered by users.
In this recipe, we will group movies based on their similarity and identify the common themes of the movies. Additionally, we will also look into how we can increase the variety of recommendations that can be provided to a user. Having said that, it is highly likely that this strategy will work less in the specific case of movie recommendations, as the variety would be much lower when compared to a retail/e-commerce...
