8. Recommender Systems

Question 1

What is the utility function for the formal model?

Answer 1

u: X x S -> R
Where X is a set of customers, S is a set of items, and R is a rating.
Essentially, we get a rating for each customer/item pairing

Question 2

What are the key problems associated with the formal model for recommender systems?

Answer 2

1. Gathering known ratings to fill the utility matrix
2. Extrapolating unknown ratings from known ones
3. Evaluating extrapolation methods in terms of success or performance

Question 3

What are the 2 ways we can collect ratings for the utility matrix?

Answer 3

1. Explicitly asking people to rate items which doesn’t work well in practice
2. Implicitly learning ratings from user actions

Question 4

What are the approaches to recommender systems for extrapolating utilities?

Answer 4

1. Content-based
2. Collaborative
3. Latent factor based

Question 5

Why is extrapolating utilities a problem?

Answer 5

Most people have not rated most items
New items have no ratings
New users have no history
Not much info to extrapolate from

Question 6

What is the main idea behind a content-based recommendation system?

Answer 6

To recommend items to customer x that are similar to previous items rated highly by that customer

Question 7

What is an item profile?

Answer 7

A set of features. It is convenient to think of it has a vector with one dimension per feature

Question 8

What is the prediction heuristic for content-based recommendation systems?

Answer 8

Given a user profile x and item profile i, estimate u(x, i) using cosine similarity between x and i

Question 9

What is a user profile and how can we calculate it?

Answer 9

When a user has rated items each with their own profile, we create a user profile using the weighted average of rated item profiles or we can weight them by the difference from the average rating for that item

Question 10

What are the pros of the content-based recommendation system?

Answer 10

1. No need for data on other users
2. Able to recommend to users with unique tastes
3. Able to recommend new and unpopular items
4. Able to provide explanations of recommendations by listing the content features that caused it to be selected

Question 11

What are the cons of the content-based recommendation system?

Answer 11

1. Finding the appropriate features is hard
2. Recommendations for new users is difficult
3. May be overspecialization where it never recommends items outside of the content profile
4. Unable to exploit quality judgements from other users

Question 12

What is the goal of a collaborative filtering system?

Answer 12

Finding a set N of other users whose ratings are similar to user x’s ratings. We estimate x’s ratings based on ratings of users in N

Question 13

What is the formula for Jaccard Similarity when we have sets of ratings for users A and B?

Answer 13

Sim(A, B) = |rA INTERSECT rB|/|rA U rB|

Question 14

What is the formula for cosine similarity when we have sets of ratings for users A and B?

Answer 14

Sim(A, B) = cos(rA, rB) = (rA * rB)/(|rA||rB|)

Question 15

What is centered cosine similarity (Pearson Correlation)?

Answer 15

Same as cosine similarity but we first normalize all ratings by subtracting the mean of the row (mean of a user’s ratings)

Question 16

What is the issue with the Jaccard and cosine similarity measures?

Answer 16

Jaccard ignores the value of the rating and only looks at overlapping things

Cosine treats missing ratings as negative by giving them a 0

Question 17

How do we translate a similarity metric to a recommendation?

Answer 17

rxi = (sim(x, y) * ryi for all y in N)/(sim(x, y) for all y in N)

Where N is the set of k users most similar to x who have rated i, rx is the vector of user x’s ratings

Question 18

What is item-item collaborative filtering?

Answer 18

Unlike user-user filtering where we compare user preferences, we want to find similar items to a given item

Question 19

What is the process for item-item collaborative filtering?

Answer 19

1. For item i, find other similar items
2. Estimate rating for item i based on ratings for similar items
3. We can use the same similarity metrics and prediction functions as the user-user model

Question 20

What is the upside of the collaborative filtering system?

Answer 20

It works for any kind of item, no feature selection is needed

Question 21

What are the cons of the collaborative filtering system?

Answer 21

1. Cold start problem where there are not enough users in the system to find a match
2. Sparisty of the user/ratings matrix means it is hard to find users that have rated the same items
   3.First rater problem where we cannot recommend an item that hasn’t been rated before
3. Popularity bias meaning we cannot recommend items to someone with a unique taste

Question 22

How do we compute a global baseline estimate and why would we need one?

Answer 22

Average rating + (movie rating - average) + (user rating - average)

We need this in case someone has not rated any movie similar to one we are trying to estimate a rating for