User profile based recommendations
Now that you have built the user profile based on the aggregate of the individual movies they enjoyed, you can compare it to the larger tfidf_summary_df DataFrame that you have been working with to generate suggestions. As you would not want to suggest movies that the user has already watched, you will first find a subset of the tfidf_summary_df DataFrame that does not contain any of the previously watched movies.
The DataFrame user_prof that you generated in the last exercise that contains a single column representing the user has been loaded for you. Similarly, the list_of_movies_enjoyed has been loaded so you can exclude them from the predictions.
Cet exercice fait partie du cours
Building Recommendation Engines in Python
Exercice interactif pratique
Essayez cet exercice en complétant cet exemple de code.
from sklearn.metrics.pairwise import cosine_similarity
# Find subset of tfidf_df that does not include movies in list_of_movies_enjoyed
tfidf_subset_df = tfidf_df.____(list_of_movies_enjoyed, axis=____)