Advertisement

Information Retrieval

, Volume 4, Issue 2, pp 133–151 | Cite as

Eigentaste: A Constant Time Collaborative Filtering Algorithm

  • Ken Goldberg
  • Theresa Roeder
  • Dhruv Gupta
  • Chris Perkins
Article

Abstract

Eigentaste is a collaborative filtering algorithm that uses universal queries to elicit real-valued user ratings on a common set of items and applies principal component analysis (PCA) to the resulting dense subset of the ratings matrix. PCA facilitates dimensionality reduction for offline clustering of users and rapid computation of recommendations. For a database of n users, standard nearest-neighbor techniques require O(n) processing time to compute recommendations, whereas Eigentaste requires O(1) (constant) time. We compare Eigentaste to alternative algorithms using data from Jester, an online joke recommending system.

Jester has collected approximately 2,500,000 ratings from 57,000 users. We use the Normalized Mean Absolute Error (NMAE) measure to compare performance of different algorithms. In the Appendix we use Uniform and Normal distribution models to derive analytic estimates of NMAE when predictions are random. On the Jester dataset, Eigentaste computes recommendations two orders of magnitude faster with no loss of accuracy. Jester is online at: http://eigentaste.berkeley.edu

recommender systems collaborative filtering dimensionality reduction jokes 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Albaum G, Best R and Hawkins D (1981) Continuous vs discrete semantic differential ratings scales. Psychological Reports, 49: 90-97.Google Scholar
  2. Arrow KJ (1963) Social Choice and Individual Values, 2nd ed. Yale University Press.Google Scholar
  3. Billsus D and Pazzani M (1998) Learning collaborative information filters. Machine Learning. In: Proceedings of the Fifteenth International Conference (ICML'98), Madison, WI, USA,24-27 July 1998. Morgan Kaufmann Publishers, San Francisco, pp. 46-54.Google Scholar
  4. Breese, Heckermen and Kadie (1998) Empirical analysis of predictive algorithms for collaborative filtering. Microsoft Research Technical Report, (MSR-TR-98-12).Google Scholar
  5. Chislenko A et al. (2000) US Patent 6092049: Method and apparatus for efficiently recommending items using automated collaborative filtering and feature-guided automated collaborative filtering.Google Scholar
  6. Dasarathy BV (1991) NN Pattern Classification Techniques. IEEE Computer Society Press, CA.Google Scholar
  7. Deerwester S, Dumais S, Furnas G, Landauer T and Harshman R (1990) Indexing by latent semantic analysis. Journal of the American Society for Information Science, 41(6): 391-407.Google Scholar
  8. Delgado JA (2000) Agent-Based Information Filtering and Recommender Systems on the Internet. PhD Thesis, Nagoya Institute of Technology.Google Scholar
  9. Ding CHQ (1999) A similarity-based probability model for latent semantic indexing. In: Proceedings of 22nd International ACM SIGIR Conference on Research and Development in Information Retrieval, Berkeley, CA, USA, August 1999. ACM, New York, NY, pp. 58-65.Google Scholar
  10. Goldberg D, Nichols D, Oki B and Terry D (1992) Using collaborative filtering to weave an information tapestry. Communication of the ACM, 35(12): 61-70.Google Scholar
  11. Gupta D and Goldberg K (1999) Jester 2.0: A linear time collaborative filtering algorithm applied to jokes. In: Proceedings of 22nd International ACM SIGIR Conference on Research and Development in Information Retrieval, Berkeley, CA, USA, August 1999. ACM, New York, NY, pp. 291-292.Google Scholar
  12. Herlocker J, Konstan J, Borchers A and Riedl J (1999) An algorithmic framework for performing collaborative filtering. In: Proceedings of 22nd International ACM SIGIR Conference on Research and Development in Information Retrieval, Berkeley, CA, USA, August 1999. ACM, New York, NY, pp. 230-237.Google Scholar
  13. Herlocker J, Konstan J, and Riedl J (2000) Explaining collaborative filtering recommendations. In: Proceedings of the ACM 2000 Conference on Computer Supported Cooperative Work, December 2-6, 2000.Google Scholar
  14. Hey J (1989) US Patent 4870579: System and method of predicting subjective reactions.Google Scholar
  15. Hofmann T (1999) Probabilistic latent semantic indexing. In: Proceedings of 22nd International ACM SIGIR Conference on Research and Development in Information Retrieval, Berkeley, CA, USA, August 1999. ACM, New York, NY, pp. 50-57.Google Scholar
  16. Hotelling H (1933) Analysis of a complex of statistical variables into principal components. Journal of Eduational Psychology, 24: 417-441.Google Scholar
  17. Jackson JE (1991) A User Guide to Principal Components: A Problem Solving Approach. John Wiley and Sons, New York.Google Scholar
  18. Konstan JA and Bharat K (1996) Integrated personal and community recommendations in collaborative filtering. CSCW Workshop.Google Scholar
  19. Konstan J, Miller B, Maltz D, Herlacker J, Gordon L and Riedl J (1997) Grouplens: Applying collaborative filtering to usenet news. Communications of the ACM, 40(3): 77-87.Google Scholar
  20. Krishnaiah PR and Kanal LN (1982) Classification, pattern recognition and reduction in dimensionality. In: Handbook of Statistics 2. North-Holland Publishing Company, Amsterdam, New York, Oxford.Google Scholar
  21. Landauer T, Littman M and Bell Communications Research (Bellcore) (1994) US Patent 5301109: Computerized cross-language document retrieval using latent semantic indexing.Google Scholar
  22. Maltz DA and Ehlrich K (1995) Pointing the way: Active collaborative filtering. Chi'95 Proceedings Papers.Google Scholar
  23. Nilsen D, ed. International Journal of Humor Research. Walter de Gruyter Inc., New York. (2000).Google Scholar
  24. Pearson K (1901) On lines and planes of closest fit to systems of points in space. Phil. Mag., 2: 559-572.Google Scholar
  25. Pennock DM and Horvitz E (1999a) Analysis of the axiomatic foundations of collaborative filtering. In: AAAI Workshop on Artificial Intelligence for Electronic Commerce, Orlando, Florida, National Conference on Arti-ficial Intelligence.Google Scholar
  26. Pennock DM and Horvitz E (1999b) Collaborative filtering by personality diagnosis:Ahybrid memory-and modelbased approach. In: IJCAI Workshop on Machine Learning for Information Filtering, Stockholm, Sweden, International Joint Conference on Artificial Intelligence.Google Scholar
  27. Pryor M (1998) The effects of singular value decomposition on collaborative filtering. Dartmouth College CS Technical Report, PCS-TR 98-338.Google Scholar
  28. Resnick P, Iacovou N, Suchak M, Bergstrom P and Riedl J (1994) Grouplens: An open architecture for collaborative filtering of netnews. Proceedings of the ACM Conference on Computer Supported Cooperative Work.Google Scholar
  29. Rich E (1979) User Modeling via Stereotypes. Cognitive Science, 3: 335-366.Google Scholar
  30. Shardanand U and Maes P (1995) Social information filtering: Algorithms for automating word of mouth. ACM Conference on Computer Human Interaction (CHI).Google Scholar
  31. Sarwar BM, Karypis G, Konstan J and Riedl JT (2000) Application of dimensionality reduction in recommender systems-a case study. ACM Conference on E-Commerce.Google Scholar
  32. Varian H and Resnick P (1997) Special issue on cf and recommender systems Communication of the ACM, 40(3).Google Scholar
  33. Ziv A (1984) Personality and Sense of Humor. Springer Publishing Co., New York.Google Scholar

Copyright information

© Kluwer Academic Publishers 2001

Authors and Affiliations

  • Ken Goldberg
    • 1
  • Theresa Roeder
    • 2
  • Dhruv Gupta
    • 2
  • Chris Perkins
    • 2
  1. 1.IEOR and EECS DepartmentsUniversity of CaliforniaBerkeleyUSA
  2. 2.IEOR DepartmentUniversity of CaliforniaBerkeleyUSA

Personalised recommendations