Recommendations from binary data
So you see from the ROEM, these models can still generate meaningful test predictions. Let's look at the actual recommendations now.
The col
function from the pyspark.sql.functions
class has been imported for you.
This exercise is part of the course
Building Recommendation Engines with PySpark
Exercise instructions
- The dataframe
original_ratings
is provided for you. It contains the original ratings that users provided. Use.filter()
to examine user 26's original ratings. - The dataframe
binary_recs
is also provided. It contains the recommendations for each user. Use.filter()
to examineuserId == 26
recommendations. Do they seem consistent with the recommendations the model provided? - Do the same thing with user 99. Feel free to examine other users.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# View user 26's original ratings
print ("User 26 Original Ratings:")
____.____(col("userId") == 26).show()
# View user 26's recommendations
print ("User 26 Recommendations:")
binary_recs.filter(col("____") == 26).show()
# View user 99's original ratings
print ("User 99 Original Ratings:")
____.filter(____("userId") == 99).show()
# View user 99's recommendations
print ("User 99 Recommendations:")
binary_recs.filter(col("userId") == ____).show()