Getting KNN data in shape
Now that you understand the ins and outs of how K-nearest neighbors works, you can leverage scikit-learn's implementation of KNN while recognizing what it is doing underneath the hood.
In the next two exercises, you will step through how to prepare your data for scikit-learn's KNN model, and then use it to make inferences about what rating a user might give a movie they haven't seen.
For consistency, you will once again be working with User_1
and the rating they would give Apollo 13 (1995)
if they saw it.
The users_to_ratings
DataFrame has again been loaded for you. This contains each user with its own row and each rating they gave as the values.
Similarly, user_ratings_table
has been loaded, which contains the raw rating values (pre-centering and filling with zeros).
This exercise is part of the course
Building Recommendation Engines in Python
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Drop the column you are trying to predict
users_to_ratings.____("Apollo 13 (1995)", axis=1, inplace=____)
# Get the data for the user you are predicting for
target_user_x = ____.____[[____]]