Knowledge and Information Systems

, Volume 55, Issue 2, pp 275–304 | Cite as

Forgetting techniques for stream-based matrix factorization in recommender systems

  • Pawel MatuszykEmail author
  • João Vinagre
  • Myra Spiliopoulou
  • Alípio Mário Jorge
  • João Gama
Regular Paper


Forgetting is often considered a malfunction of intelligent agents; however, in a changing world forgetting has an essential advantage. It provides means of adaptation to changes by removing effects of obsolete (not necessarily old) information from models. This also applies to intelligent systems, such as recommender systems, which learn users’ preferences and predict future items of interest. In this work, we present unsupervised forgetting techniques that make recommender systems adapt to changes of users’ preferences over time. We propose eleven techniques that select obsolete information and three algorithms that enforce the forgetting in different ways. In our evaluation on real-world datasets, we show that forgetting obsolete information significantly improves predictive power of recommender systems.


Recommender systems Forgetting techniques Matrix factorization Data stream mining Machine learning 



We would like to thank to the Institute of Psychology II at the University of Magdeburg for making their computational cluster available for our experiments. This work is financed by the ERDF European Regional Development Fund through the Operational Programme for Competitiveness and Internationalisation—COMPETE 2020 Programme, and by National Funds through the FCT Fundação para a Ciência e a Tecnologia (Portuguese Foundation for Science and Technology) within project POCI-01-0145-FEDER-006961.


  1. 1.
    Babcock B, Babu S, Datar M, Motwani R, Widom J (2002) Models and issues in data stream systems. In: Lucian P, Serge A, Phokion GK (eds) PODS. ACM, pp 1–16Google Scholar
  2. 2.
    Berry MW, Dumais ST, O’Brien GW (1995) Using linear algebra for intelligent information retrieval. SIAM Review 37(4):573–595MathSciNetCrossRefzbMATHGoogle Scholar
  3. 3.
    Cao F, Ester M, Qian W, Zhou A (2006) Density-based clustering over an evolving data stream with noise. In: Joydeep G, Diane L, David BS, Jaideep S (eds) SDM. SIAMGoogle Scholar
  4. 4.
    Celma O (2010) Music recommendation and discovery in the long tail. Springer, BerlinCrossRefGoogle Scholar
  5. 5.
    Chua FCT, Oentaryo RJ, Lim E-P (2013) Modeling temporal adoptions using dynamic matrix factorization. In: Hui X, George K, Bhavani MT, Diane JC, Xindong W (eds) ICDM, IEEE computer society, pp 91–100Google Scholar
  6. 6.
    Cremonesi P, Koren Y, Turrin R (2010) Performance of recommender algorithms on top-n recommendation tasks. In: Proceedings of ACM RecSys, RecSys ’10. ACM, pp 39–46Google Scholar
  7. 7.
    Ding Y, Li X (2005) Time weight collaborative filtering. In: Otthein H, Hans-Jörg S, Norbert F, Abdur C, Wilfried T, (eds) CIKM. ACM, pp 485–492Google Scholar
  8. 8.
    Domingos P, Hulten G (2001) Catching up with the data: research issues in mining data streams. In: DMKDGoogle Scholar
  9. 9.
    Gama João (2012) A survey on learning from data streams: current and future trends. Prog Artif Intell 1(1):45–55CrossRefGoogle Scholar
  10. 10.
    Gama J, Sebastião R, Rodrigues PP (2009) Issues in evaluation of stream learning algorithms. In: KDDGoogle Scholar
  11. 11.
    Halchenko YO, Hanke M (2012) Open is not enough. Let’s take the next step: an integrated, community-driven computing platform for neuroscience. Front NeuroinfGoogle Scholar
  12. 12.
    Harper F Maxwell, Konstan Joseph A (2016) The movielens datasets: history and context. TiiS 5(4):19Google Scholar
  13. 13.
    Hu Y, Koren Y, Volinsky C (2008) Collaborative filtering for implicit feedback datasets. In: Proceedings of the 8th IEEE international conference on data mining (ICDM 2008), December 15–19, 2008, Pisa, Italy. IEEE Computer Society, pp 263–272Google Scholar
  14. 14.
    Hulten G, Spencer L, Domingos P (2001) Mining time-changing data streams. ACM SIGKDDGoogle Scholar
  15. 15.
    Hulten G, Spencer L, Domingos P (2001) Mining time-changing data streams. In: Doheon L, Mario S, Foster JP, Ramakrishnan S (eds) KDD. ACM, pp 97–106Google Scholar
  16. 16.
    Karimi R, Freudenthaler C, Nanopoulos A, Schmidt-Thieme L (2011) Towards optimal active learning for matrix factorization in recommender systems. In: ICTAI. IEEE, pp 1069–1076Google Scholar
  17. 17.
    Koren Y, Bell R, Volinsky C (2009) Matrix factorization techniques for recommender systems. Computer 42(8):30–37CrossRefGoogle Scholar
  18. 18.
    Koren Y (2009) Collaborative filtering with temporal dynamics. In: KDDGoogle Scholar
  19. 19.
    Koychev I (2000) Gradual forgetting for adaptation to concept drift. In: ECAI 2000 workshop on current issues in spatio-temporal reasoning, Berlin, Germany, pp 101–106Google Scholar
  20. 20.
    Koychev I, Schwab I (2000) Adapting to drifting user’s interests. In: ECML 2000Google Scholar
  21. 21.
    Li Xue, Barajas Jorge M, Ding Yi (2007) Collaborative filtering on streaming data with interest-drifting. Intell Data Anal 11(1):75–87Google Scholar
  22. 22.
    Ling G, Yang H, King I, Lyu MR (2012) Online learning for collaborative filtering. In: International joint conference on neural networks. IEEE, pp 1–8Google Scholar
  23. 23.
    Liu NN, Zhao M, Xiang EW, Yang Q (2010) Online evolutionary collaborative filtering. In: Proceedings of the ACM RecSysGoogle Scholar
  24. 24.
    Massa P, Avesani P (2006) Trust-aware bootstrapping of recommender systems. In: ECAI workshop on recommender systems. Citeseer, pp 29–33Google Scholar
  25. 25.
    Matuszyk P, Spiliopoulou M (2014) Selective forgetting for incremental matrix factorization in recommender systems. In: Discovery science, volume 8777 of LNCS. Springer, pp 204–215Google Scholar
  26. 26.
    Matuszyk P, Spiliopoulou M (2015) Semi-supervised learning for stream recommender systems. In: Nathalie J, Stan M (eds) Discovery science, volume 9356 of LNCS. Springer, pp 131–145Google Scholar
  27. 27.
    Matuszyk P, Vinagre J, Spiliopoulou M, Jorge AM, Gama J (2015) Forgetting methods for incremental matrix factorization in recommender systems. In: Proceedings of the ACM SAC, SAC ’15, New York, NY, USA. ACM, pp 947–953Google Scholar
  28. 28.
    Miranda C, Jorge AM (2008) Incremental collaborative filtering for binary ratings. In: Web intelligence conference proceedings. IEEE Computer Society, pp 389–392Google Scholar
  29. 29.
    Nasraoui O, Cerwinske J, Rojas C, González F (2007) Performance of recommendation systems in dynamic streaming environments. In: SDM, SIAMGoogle Scholar
  30. 30.
    Nasraoui O, Uribe CC, Coronel CR, González FA (2003) TECNO-STREAMS: tracking evolving clusters in noisy data streams with a scalable immune system learning model. In: Proceedings of the IEEE ICDM 2003, pp 235–242Google Scholar
  31. 31.
    Papagelis M, Rousidis I, Plexousakis D, Theoharopoulos E (2005) Incremental collaborative filtering for highly-scalable recommendation algorithms. In: Proceedings of the ISMIS 2005Google Scholar
  32. 32.
    Paterek A (2007) Improving regularized singular value decomposition for collaborative filtering. In: Proceedings of the KDD cup workshop at SIGKDD’07Google Scholar
  33. 33.
    Sarwar BM, Karypis G, Konstan JA, Riedl JT (2000) Application of dimensionality reduction in recommender system—a case study. In: In ACM WEBKDD workshopGoogle Scholar
  34. 34.
    Shaffer JP (1995) Multiple hypothesis testing. Ann Rev Psychol 46(1):561–584CrossRefGoogle Scholar
  35. 35.
    Sun John Z, Parthasarathy Dhruv, Varshney Kush R (2014) Collaborative Kalman filtering for dynamic matrix factorization. IEEE Trans Signal Process 62(14):3499–3509MathSciNetCrossRefGoogle Scholar
  36. 36.
    Takács G, Pilászy I, Németh B, Tikk D (2009) Scalable collaborative filtering approaches for large recommender systems. J Mach Learn Res 10:623–656Google Scholar
  37. 37.
    Joáo V, Jorge AM (2012) Forgetting mechanisms for scalable collaborative filtering. J Braz Comput Soc 18(4):271–282CrossRefGoogle Scholar
  38. 38.
    Vinagre J, Jorge AM, Gama J (2014) Fast incremental matrix factorization for recommendation with positive-only feedback. In: UMAP, pp 459–470Google Scholar

Copyright information

© Springer-Verlag London Ltd. 2017

Authors and Affiliations

  • Pawel Matuszyk
    • 1
    Email author
  • João Vinagre
    • 2
  • Myra Spiliopoulou
    • 1
  • Alípio Mário Jorge
    • 2
  • João Gama
    • 3
  1. 1.Knowledge Management and Discovery LabOtto von Guericke UniversityMagdeburgGermany
  2. 2.LIAAD - INESC TECFCUP - Universidade do PortoPortoPortugal
  3. 3.LIAAD - INESC TECFEP - Universidade do PortoPortoPortugal

Personalised recommendations