31. [Exercise] Improve the Quality of Similiar Movies
LEARNING ACTIVITY
Exercise
Improve the results! Some ideas to try:
Discard bad ratings - only recommend good movies
Try different similarity metrics (Pearson Correlation Coefficient, Jaccard Coefficient, Conditional Probability)
Adjust the thresholds for minimum co-raters or minimum score
Invent a new similarity metric that takes the number of co-raters into account
Use genre information in u.items to boost scores from movies in the same genre
Previous30. [Activity] Running the Similiar Movies Script using Spark's Cluster ManagerNext32. [Activity] Using spark-submit to run Spark driver scripts
Last updated