Skip to main content
SpringerLink
Log in

Graph bandit for diverse user coverage in online recommendation

  • Published:
Applied Intelligence Aims and scope Submit manuscript

Abstract

We study a recommendation system problem, in which the system must be able to cover as many users’ preferences as possible while these preferences change over time. This problem can be formulated as a variation of the maximum coverage problem; specifically we introduced a novel problem of Online k-Hitting Set, where the number of sets and elements within the sets can change dynamically. When the number of distinctive elements is large, an exhaustive search for even a fixed number of elements is known to be computationally expensive. Even the static problem is known to be NP-hard (Hochba, ACM SIGACT News 28(2):40–52, 1997) and many known algorithms tend to have exponential growth in complexity. We propose a novel graph based UCB1 algorithm that effectively minimizes the number of elements to consider, thereby reducing the search space greatly. The algorithm utilizes a new rewarding scheme to choose items that satisfy more users by balancing coverage and diversity as it construct a relational graph between items to recommend. Experiments show that the new graph based algorithm performs better than existing techniques such as Ranked Bandit (Radlinski et al. 2008) and Independent Bandits (Kohli et al. 2013) in terms of satisfying diverse types of users while minimizing computational complexity.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11

Similar content being viewed by others

References

  1. Agrawal S (2011) Optimization under uncertainty: bounding the correlation gap. Ph D Thesis. Stanford University

  2. Auer P, Cesa-Bianchi N, Fischer P (2002) Finite-time analysis of the multiarmed bandit problem. Mach Learn 47(2-3):235–256

    Article  MATH  Google Scholar 

  3. Ausiello G, Boria N, Giannakos A, Lucarelli G, Paschos VT (2011) Online maximum k-coverage. In: International symposium on fundamentals of computation theory. Springer, pp 181–192

  4. Azizi N, Zolfaghari S (2004) Adaptive temperature control for simulated annealing: a comparative study. Comput Oper Res 31(14):2439–2451

    Article  MathSciNet  MATH  Google Scholar 

  5. Bnaya Z, Puzis R, Stern R, Felner A (2013) Volatile multi-armed bandits for guaranteed targeted social crawling. In: Late-breaking developments in the field of artificial intelligence. Bellevue, Washington, USA, p 2013

  6. Burke R (2007) The adaptive web. In: Brusilovsky P, Kobsa A, Nejdl W (eds) Hybrid web recommender systems. Springer, Berlin, Heidelberg, pp 377–408

    Google Scholar 

  7. Chakrabarti D, Kumar R, Radlinski F, Upfal E (2008) Mortal multi-armed bandits. In: Advances in neural information processing systems 21, proceedings of the 22nd annual conference on neural information processing systems. Vancouver, British Columbia, Canada, pp 273–280

  8. Cohen R, Katzir L (2008) The generalized maximum coverage problem. Inf Process Lett 108(1):15–22

    Article  MathSciNet  MATH  Google Scholar 

  9. Diriye A, White R, Buscher G, Dumais S (2012) Leaving so soon?: understanding and predicting web search abandonment rationales. In: Proceedings of the 21st ACM international conference on information and knowledge management. ACM, CIKM’12, New York, pp 1025–1034

  10. Ekstrand MD, Riedl JT, Konstan JA (2011) Collaborative filtering recommender systems. Foundation and Trends in Human Computer Interaction 4(2):81–173

    Article  Google Scholar 

  11. Goldberg K, Roeder T, Gupta D, Perkins C (2001) Eigentaste: a constant time collaborative filtering algorithm. Inf Retr 4(2):133–151

    Article  MATH  Google Scholar 

  12. Harper FM, Konstan JA (2015) The movielens datasets: history and context. ACM Trans on Interactactive Intell Syst 5(4):19:1–19:19

    Google Scholar 

  13. Hochba DS (1997) Approximation algorithms for np-hard problems. ACM SIGACT News 28(2):40–52

    Article  Google Scholar 

  14. Hochbaum DS, Pathria A (1998) Analysis of the greedy approach in problems of maximum k-coverage. Nav Res Logista (NRL) 45(6):615–627

    Article  MathSciNet  MATH  Google Scholar 

  15. Joachims T (2002) Optimizing search engines using clickthrough data. In: Proceedings of the eighth ACM SIGKDD international conference on knowledge discovery and data mining. ACM, KDD’02, New York, , pp 133–142

  16. Kohli P, Salek M, Stoddard G (2013) A fast bandit algorithm for recommendations to users with heterogeneous tastes. In: Proceedings of the 27th AAAI conference on artificial intelligence. AAAI AAAI’13, Press, pp 1135–1141

  17. Lee K, Lee K (2015) Escaping your comfort zone: a graph-based recommender system for finding novel recommendations among relevant items. Expert Syst Appl 42(10):4851–4858

    Article  Google Scholar 

  18. Pandey S, Chakrabarti D, Agarwal D (2007) Multi-armed bandit problems with dependent arms. In: Proceedings of the 24th international conference on machine learning. ACM, ICML’07, New York, pp 721–728

  19. Park W, Oh JC, Blowers MK, Wolf MB (2006) An open-set speaker identification system using genetic learning classifier system. In: Proceedings of the 8th annual conference on genetic and evolutionary computation. ACM, GECCO’06, New York, pp 1597–1598

  20. Pazzani MJ, Billsus D (2007) Content-based recommendation systems. In: The adaptive web: method and strategies of web personalization. Vol 4321 of lecture notes in computer science. Springer, pp 325–341

  21. Radlinski F, Kleinberg R, Joachims T (2008) Learning diverse rankings with multi-armed bandits. In: Proceedings of the 25th international conference on machine learning. ACM, ICML’08, New York, pp 784–791

  22. Rahman M, Oh JC (2015a) Fast online learning to recommend a diverse set from big data. In: The 28th international conference on industrial, engineering and other application of applied intelligent systems IEA/AIE’15. Springer, Switzerland, pp 361–370

  23. Rahman M, Oh JC (2015b) Parallel and synchronized UCB2 for online recommendation systems. In: IEEE/WIC/ACM international conference on web intelligence and intelligent agent technology, WI-IAT 2015. Singapore, December 6–9, 2015, vol I, pp 413–416

  24. Robertson SE (1997) In: Sparck Jones K, Willett P (eds) Readings in information retrieval. Morgan Kaufmann Publishers Inc., San Francisco, CA, USA, pp 281–286

  25. Shani G, Gunawardana A (2011) Evaluating recommendation systems. In: Recommender systems handbook. Springer, pp 257–297

  26. Sviridenko M (2004) A note on maximizing a submodular set function subject to a knapsack constraint. Oper Res Lett 32(1):41–43

    Article  MathSciNet  MATH  Google Scholar 

  27. Tekin C, Liu M (2012) Online learning of rested and restless bandits. IEEE Trans Inf Theory 58(8):5588–5611

    Article  MathSciNet  MATH  Google Scholar 

  28. Vermorel J, Mohri M (2005) Multi-armed bandit algorithms and empirical evaluation. In: Proceedings of the 16th european conference on machine learning ECML’05. Springer, Berlin, Heidelberg, pp 437–448

  29. Vinterbo SA, Øhrn A (2000) Minimal approximate hitting sets and rule templates. Int J Approx Reason 25:123–143

    Article  MathSciNet  MATH  Google Scholar 

  30. Zhang M, Hurley N (2008) Avoiding monotony: improving the diversity of recommendation lists Proceedings of the 2008 ACM conference on recommender systems. ACM, RecSys’08, New York, pp 123–130

    Chapter  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. DMA Lab, EECS, Syracuse University, Syracuse, NY, USA

    Mahmuda Rahman & Jae C. Oh

Authors
  1. Mahmuda Rahman
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jae C. Oh
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Mahmuda Rahman.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Rahman, M., Oh, J.C. Graph bandit for diverse user coverage in online recommendation. Appl Intell 48, 1979–1995 (2018). https://doi.org/10.1007/s10489-017-0977-1

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10489-017-0977-1

Keywords

Search

Navigation