Scale And Translation Invariant Collaborative Filtering Systems

Abstract

Collaborative filtering systems are prediction algorithms over sparse data sets of user preferences. We modify a wide range of state-of-the-art collaborative filtering systems to make them scale and translation invariant and generally improve their accuracy without increasing their computational cost. Using the EachMovie and the Jester data sets, we show that learning-free constant time scale and translation invariant schemes outperforms other learning-free constant time schemes by at least 3% and perform as well as expensive memory-based schemes (within 4%). Over the Jester data set, we show that a scale and translation invariant Eigentaste algorithm outperforms Eigentaste 2.0 by 20%. These results suggest that scale and translation invariance is a desirable property.

Keywords

Recommender System, Incomplete Vectors, Regression, e-Commerce.

Reference

Daniel Lemire, Scale and Translation Invariant Collaborative Filtering Systems, Information Retrieval, 8 (1), pages 129-150, January 2005. (NRC 46508)

Download

• Scale And Translation Invariant Collaborative Filtering Systems (pdf)

Hint : It is sometimes necessary to hold down shift while clicking in order to save a document. The paper is also available through Cogprints.

Software

Download the java implementation of the collaborative filtering algorithms. It requires the gnu.trove package. New: COFI has been open sourced under GPL.

BibTeX

@article{LemireIR2003,
  author = 	 "Daniel Lemire",
  title = 	 "Scale and Translation Invariant Collaborative Filtering Systems",
  journal = 	 {Information Retrieval},
  volume =      "8",
  number = "1",
  pages =       "129--150",
  month= "January",
  year =         {2005}
}

Author

• Daniel Lemire: lemire at acm.org

Related work

• Daniel Lemire's publications

Citations

This paper is cited by

1. Chen-Fan Wu, A Study for the Effect of the Recommendation Strategy and Information on the on-line Recommendation Performance, M.Sc. Thesis, Chung Yuan Christian University, 2004.
2. Matthew Garden, On the use of semantic feedback in recommender systems, M.Sc. Thesis in Computer Science, McGill University, August 2004.
3. Michelle Anderson, Marcel Ball, Harold Boley, Stephen Greene, Nancy Howse, Daniel Lemire, Sean McGrath, RACOFI: Rule-Applying Collaborative Filtering Systems, Proceedings IEEE/WIC COLA'03, Halifax, Canada, October 2003. (NRC 46507)
4. J. Fiaidhi, K. Passi, and S. Mohammed, Developing a Framework for Learning Objects Search Engine, The 2004 International Conference on Internet Computing (IC04), Las Vegas, Nevada, USA, June 21-24, 2004.
5. J. Fiaidhi, RecoSearch: A Model for Collaboratively Filtering Java Learning Objects, ITDL Vol 1. No. 7., 2004. (html, pdf)
6. Jasna Tusek, Item-based Collaborative Filtering aus der Perspektive des Customer Relationship Managements, Seminararbeit aus Informationswirtschaft, Wirtschafts Universtät, May 2005.
7. Huseyin Polat and Wenliang Du, Privacy-preserving collaborative filtering, International Journal of Electronic Commerce, 9 (4), pages 9-35, 2005.
8. Xiaohua Sun, Fansheng Kong, Xiaobing Yang, Song Ye, Using Latent Class Models for Neighbors Selection in Collaborative Filtering, Lecture Notes in Computer Science, Volume 3584, 2005.