1 Recitation Problems
These problems are to be found in: Mining of Massive Datasets, Online Edition by Jure Leskovec, Anand Rajaraman, Jeff Ullman.
1.1 Chapter 9
Problems: 9.2.1,9.2.3,9.3.1,9.4.1
2 Practicum Problems
2.1 Problem 1
Load the Movielens 100k dataset (ml-100k.zip) into Python using Pandas dataframes. Build a user profile on centered data (by user rating) for both users 200 and 15, and calculate the cosine similarity and distance between the user’s preferences and the item/movie 95. Which user would a recommender system suggest this movie to?
2.2 Problem 2
Load the Movielens 100k dataset (ml-100k.zip) into Python using Pandas dataframes. Convert the ratings data into a utility matrix representation, and find the 10 most similar users for user 1 based on cosine similarity of the centered user ratings data. Based on the average of of the ratings for item 508 from the similar users, what is the expected rating for this item for user 1?
Prof. Panchal:
