The best indicator that a passenger will show up to board the flight

1. Filtering and Recommender SystemsContent-based and Collaborative 4/15 The best indicator that a passenger will show up to board the flight is that she called in for a special meal

2. Filtering and Recommender SystemsContent-based and Collaborative Some of the slides based On Mooney’s Slides

3. Personalization • Recommenders are instances of personalization software. • Personalization concerns adapting to the individual needs, interests, and preferences of each user. • Includes: • Recommending • Filtering • Predicting (e.g. form or calendar appt. completion) • From a business perspective, it is viewed as part of Customer Relationship Management (CRM).

4. Feedback & Prediction/Recommendation • Traditional IR has a single user—probably working in single-shot modes • Relevance feedback… • WEB search engines have: • Working continually • User profiling • Profile is a “model” of the user • (and also Relevance feedback) • Many users • Collaborative filtering • Propagate user preferences to other users… You know this one

5. Recommender Systems in Use • Systems for recommending items (e.g. books, movies, CD’s, web pages, newsgroup messages) to users based on examples of their preferences. • Many on-line stores provide recommendations (e.g. Amazon, CDNow). • Recommenders have been shown to substantially increase sales at on-line stores.

6. Feedback Detection Non-Intrusive Intrusive • Explicitly ask users to rate items/pages • Click certain pages in certain order while ignore most pages. • Read some clicked pages longer than some other clicked pages. • Save/print certain clicked pages. • Follow some links in clicked pages to reach more pages. • Buy items/Put them in wish-lists/Shopping Carts

7. Justifying Recommendation.. • Recommendation systems must justify their recommendations • Even if the justification is bogus.. • For search engines, the “justifications” are the page synopses • Some recommendation algorithms are better at providing human-understandable justifications than others • Content-based ones can justify in terms of classifier features.. • Collaborative ones are harder-pressed other than saying “people like you seem to like this stuff” • In general, giving good justifications is important..

8. Content-based vs. Collaborative Recommendation Needs description of items… Needs only ratings from other users

9. Content-Based Recommending • Recommendations are based on information on the content of items rather than on other users’ opinions. • Uses machine learning algorithms to induce a profile of the users preferences from examples based on a featural description of content. • Lots of systems

10. Vector of Bags model E.g. Books have several different fields that are all text Authors, description, … A word appearing in one field is different from the same word appearing in another Want to keep each bag different—vector of m Bags; Conditional probabilities for each wordw.r.teach class and bag Can give a profile of a user in terms of words that are most predictive of what they like Strenghof a keyword Log[P(w|rel)/P(w|~rel)] We can summarize a user’s profile in terms of the words that have strength above some threshold. Related to mutual information Adapting Naïve Bayes idea for Book Recommendation

11. User Database A 9 B 3 C : : Z 5 A B C 9 : : Z 10 A 5 B 3 C : : Z 7 A B C 8 : : Z A 6 B 4 C : : Z A 10 B 4 C 8 . . Z 1 A 9 B 3 C . . Z 5 A 9 B 3 C : : Z 5 A 10 B 4 C 8 . . Z 1 Correlation Match Extract Recommendations C Active User Collaborative Filtering Correlation analysis Here is similar to the Association clusters Analysis!

12. Item-User Matrix • The input to the collaborative filtering algorithm is an mxn matrix where rows are items and columns are users • Sort of like term-document matrix (items are terms and documents are users) • Can think of users as vectors in the space of items (or vice versa) • Can do vector similarity between users • And find who are most similar users.. • Can do scalar clusters over items etc.. • And find what are most correlated items Think usersdocs Itemskeywords

13. A Collaborative Filtering Method(think kNN) • Weight all users with respect to similarity with the active user. • How to measure similarity? • Could use cosine similarity; normally pearson coefficient is used • Select a subset of the users (neighbors) to use as predictors. • Normalize ratings and compute a prediction from a weighted combination of the selected neighbors’ ratings. • Present items with highest predicted ratings as recommendations.

14. Finding User Similarity with Person Correlation Coefficient • Typically use Pearson correlation coefficient between ratings for active user, a, and another user, u. ra and ru are the ratings vectors for the m items rated by botha and u ri,j is user i’s rating for item j

15. Person Correlation Coefficient is the same as vector similarity over centered ratings vectors • It is easy to check for yourself that pearson correlation coefficient is the same as the cosine theta distance between centered ratings vectors • Covariance = dot product • Sqrt (Variance of each vector) = norm of each vector

16. Neighbor Selection • For a given active user, a, select correlated users to serve as source of predictions. • Standard approach is to use the most similar k users, u, based on similarity weights, wa,u • Alternate approach is to include all users whose similarity weight is above a given threshold.

17. Rating Prediction • Predict a rating, pa,i, for each item i, for active user, a, by using the k selected neighbor users, u  {1,2,…k}. • To account for users different ratings levels, base predictions on differences from a user’s average rating. • Weight users’ ratings contribution by their similarity to the active user. ri,j is user i’s rating for item j

18. Similarity Weighting=User Similarity • Typically use Pearson correlation coefficient between ratings for active user, a, and another user, u. ra and ru are the ratings vectors for the m items rated by botha and u ri,j is user i’s rating for item j

19. Significance Weighting • Important not to trust correlations based on very few co-rated items. • Include significance weights, sa,u, based on number of co-rated items, m.

20. Item-centered Collaborative Filtering • Starting with a “centered” user-item matrix, we found k-nearest users to the active user and used them to recommend unrated items • We can also use the centered U-I matrix to compute item-item correlations by starting with U-I’xU-I, and doing (a) association clusters and (b) scalar clusters • This will give us, for each item, k-nearest items • Now, given a new item In to be rated for a user U, we first find k items closest to In and, and take their (weighted) average rating from the user U as predictive of U’s rating of In • An advantage of this method over the “user-centered” idea is that the justifications for the recommendations can be more meaningful (you can tell the user that we are recommending In because she rated the items in its association cluster high..)

21. LSI-style techniques for collaborative filtering • The NETFLIX prize was won by an approach that did “latent factor analysis” (aka LSI) on the u-i matrix, so that both users and items are seen as vectors in a k-dimensional factor space • One technical difficulty in doing LSI on u-i matrix is that it has many “null” values • D-t matrix is sparse and that is good. U-I matrix has null values and that is bad (because null != 0) • Two approaches: • “fill in” the missing ratings (“Imputation” method) so we have no more null values • “compute distance between vectors only in terms of their common non-null dimensions • Problem: Overfitting. Solution: Regularization—penalize “large factor” values. qi item in factor space pu user in factor space

22. Problems with Collaborative Filtering • Cold Start: There needs to be enough other users already in the system to find a match. • Sparsity: If there are many items to be recommended, even if there are many users, the user/ratings matrix is sparse, and it is hard to find users that have rated the same items. • First Rater: Cannot recommend an item that has not been previously rated. • New items • Esoteric items • Popularity Bias: Cannot recommend items to someone with unique tastes. • Tends to recommend popular items. • WHAT DO YOU MEAN YOU DON’T CARE FOR BRITNEY SPEARS YOU DUNDERHEAD?#$%$%$&^

23. Advantages of Content-Based Approach • No need for data on other users. • No cold-start or sparsity problems. • Able to recommend to users with unique tastes. • Able to recommend new and unpopular items • No first-rater problem. • Can provide explanations of recommended items by listing content-features that caused an item to be recommended. • Well-known technology The entire field of Classification Learning is at (y)our disposal!

24. Disadvantages of Content-Based Method • Requires content that can be encoded as meaningful features. • Users’ tastes must be represented as a learnable function of these content features. • Unable to exploit quality judgments of other users. • Unless these are somehow included in the content features.

25. User-ratings Vector Training Examples Content-Based Predictor Pseudo User-ratings Vector User-rated Items Unrated Items Items with Predicted Ratings Content-Boosted CF - I

26. Content-Boosted CF - II User Ratings Matrix Pseudo User Ratings Matrix Content-Based Predictor • Compute pseudo user ratings matrix • Full matrix – approximates actual full user ratings matrix • Perform CF • Using Pearson corr. between pseudo user-rating vectors • This works better than either!

27. Why can’t the pseudo ratings be used to help content-based filtering? • How about using the pseudo ratings to improve a content-based filter itself? (or how access to unlabelled examples improves accuracy…) • Learn a NBC classifier C0 using the few items for which we have user ratings • Use C0 to predict the ratings for the rest of the items • Loop • Learn a new classifier C1 using all the ratings (real and predicted) • Use C1 to (re)-predict the ratings for all the unknown items • Until no change in ratings • With a small change, this actually works in finding a better classifier! • Change: Keep the class posterior prediction (rather than just the max class) • This means that each (unlabelled) entity could belong to multiple classes—with fractional membership in each • We weight the counts by the membership fractions • E.g. P(A=v|c) = Sum of class weights of all examples in c that have A=v divided by Sum of class weights of all examples in c • This is called expectation maximization • Very useful on web where you have tons of data, but very little of it is labelled • Reminds you of K-means, doesn’t it? • (no coincidence—K-means is “hard-assignment” EM)

29. (boosted) content filtering

30. You train me—I train you… Co-training Small labeled data needed • Suppose each instance has two parts: x = [x1, x2] x1, x2 conditionally independent given f(x) • Suppose each half can be used to classify instance f1, f2 such that f1(x1) = f2(x2) = f(x) • Suppose f1, f2 are learnable f1  H1, f2  H2,  learning algorithms A1, A2 ~ A2 A1 [x1, x2] <[x1, x2], f1(x1)> f2 Unlabeled Instances Labeled Instances Hypothesis

31. Learning to classify web pages as course pages x1 = bag of words on a page x2 = bag of words from all anchors pointing to a page Naïve Bayes classifiers 12 labeled pages 1039 unlabeled It really works!

32. Observations • Can apply A1 to generate as much training data as one wants • If x1 is conditionally independent of x2 / f(x), • then the error in the labels produced by A1 • will look like random noise to A2 !!! • Thus no limit to quality of the hypothesis A2 can make

34. Discussion of the Google News Collaborative Filtering Paper