Data Mining and Knowledge Discovery

, Volume 30, Issue 4, pp 964–994 | Cite as

Characterizing concept drift

  • Geoffrey I. WebbEmail author
  • Roy Hyde
  • Hong Cao
  • Hai Long Nguyen
  • Francois Petitjean


Most machine learning models are static, but the world is dynamic, and increasing online deployment of learned models gives increasing urgency to the development of efficient and effective mechanisms to address learning in the context of non-stationary distributions, or as it is commonly called concept drift. However, the key issue of characterizing the different types of drift that can occur has not previously been subjected to rigorous definition and analysis. In particular, while some qualitative drift categorizations have been proposed, few have been formally defined, and the quantitative descriptions required for precise and objective understanding of learner performance have not existed. We present the first comprehensive framework for quantitative analysis of drift. This supports the development of the first comprehensive set of formal definitions of types of concept drift. The formal definitions clarify ambiguities and identify gaps in previous definitions, giving rise to a new comprehensive taxonomy of concept drift types and a solid foundation for research into mechanisms to detect and address concept drift.


Concept drift Learning from non-stationary distributions Stream learning Stream mining 



We are grateful to David Albrecht, Mark Carman, Bart Goethals, Nayyar Zaidi and the anonymous reviewers for valuable comments and suggestions. This research has been supported by the Australian Research Council under grant DP140100087 and Asian Office of Aerospace Research and Development, Air Force Office of Scientific Research under contract FA2386-15-1-4007.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.


  1. Aggarwal CC (2009) Data streams: an overview and scientific applications. Springer, Berlin, pp 377–397. doi: 10.1007/978-3-642-02788-8_14
  2. Aggarwal CC, Han J, Wang J, Yu PS (2003) A framework for clustering evolving data streams. In: Proceedings of the 29th international conference on very large data bases, VLDB Endowment, 29:81–92Google Scholar
  3. Angluin D (1988) Queries and concept learning. Mach Learn 2(4):319–342Google Scholar
  4. Babcock B, Datar M, Motwani R (2002) Sampling from a moving window over streaming data. In: Proceedings of the thirteenth annual ACM-SIAM symposium on discrete algorithms, Society for Industrial and Applied Mathematics, pp 633–634Google Scholar
  5. Baena-Garcıa M, del Campo-Ávila J, Fidalgo R, Bifet A, Gavalda R, Morales-Bueno R (2006) Early drift detection method. In: Fourth international workshop on knowledge discovery from data streams, 6:77–86Google Scholar
  6. Bartlett PL, Ben-David S, Kulkarni SR (2000) Learning changing concepts by exploiting the structure of change. Mach Learn 41(2):153–174CrossRefzbMATHGoogle Scholar
  7. Bifet A, Gama J, Pechenizkiy M, Zliobaite I (2011) Handling concept drift: importance, challenges and solutions. PAKDD-2011 Tutorial, Shenzhen, ChinaGoogle Scholar
  8. Bifet A, Gavaldà R (2009) Adaptive learning from evolving data streams. In: Advances in intelligent data analysis VIII, Springer, 249–260Google Scholar
  9. Bifet A, Holmes G, Kirkby R, Pfahringer B (2010a) MOA: massive online analysis. J Mach Learn Res 11:1601–1604Google Scholar
  10. Bifet A, Holmes G, Pfahringer B (2010b) Leveraging bagging for evolving data streams. In: Machine learning and knowledge discovery in databases, Springer, pp 135–150Google Scholar
  11. Bose RJC, van der Aalst WMP, Zliobaite I, Pechenizkiy M (2011) Handling concept drift in process mining. In: Haralambos M, Colette R (eds) Advanced information systems engineering., Lecture notes in computer science, Springer, Berlin, pp 391–405. doi: 10.1007/978-3-642-21640-4_30
  12. Brzezinski D (2014a) Reacting to different types of concept drift: the accuracy updated ensemble algorithm. Neural Netw Learn Syst IEEE Trans 25(1):81–94. doi: 10.1109/TNNLS.2013.2251352 CrossRefGoogle Scholar
  13. Brzeziński D (2010) Mining data streams with concept drift. Master’s thesis, Poznan University of TechnologyGoogle Scholar
  14. Brzezinski D, Stefanowski J (2014b) Reacting to different types of concept drift: the accuracy updated ensemble algorithm. Neural Netw Learn Syst IEEE Trans 25(1):81–94CrossRefGoogle Scholar
  15. Brzezinski D, Stefanowski J (2014c) Prequential AUC for classifier evaluation and drift detection in evolving data streams. In: Proceedings of the 3rd international workshop on new frontiers in mining complex patterns, NancyGoogle Scholar
  16. Cieslak DA, Chawla NV (2009) A framework for monitoring classifiers performance: when and why failure occurs? Knowl Inform Syst 18(1):83–108 ISSN 0219-1377CrossRefGoogle Scholar
  17. Dongre PB, Malik LG (2014) A review on real time data stream classification and adapting to various concept drift scenarios. In: Advance computing conference (IACC), 2014 IEEE international, pp 533–537, doi: 10.1109/IAdCC.2014.6779381
  18. Dries Anton, Rückert Ulrich (2009) Adaptive concept drift detection. Stat Anal Data Min 2(5–6):311–327MathSciNetCrossRefGoogle Scholar
  19. Gaber Mohamed Medhat, Zaslavsky Arkady, Krishnaswamy Shonali (2005) Mining data streams: a review. ACM Sigmod Rec 34(2):18–26CrossRefzbMATHGoogle Scholar
  20. Gama J, Zliobaite I, Bifet A, Pechenizkiy M, Bouchachia A (2014) A survey on concept drift adaptation. ACM Comput Surv 46(4):44:1–44:37. doi: 10.1145/2523813 ISSN 0360–0300CrossRefzbMATHGoogle Scholar
  21. Gama J, Rodrigues P (2009) An overview on mining data streams, volume 206 of studies in computational intelligence. Springer, Berlin. doi: 10.1007/978-3-642-01091-0_2 Google Scholar
  22. Gama J, Medas P, Castillo G, Rodrigues P (2004) Learning with drift detection. In Ana LC, Bazzan, Sofiane L (ed), Advances in artificial intelligence SBIAGoogle Scholar
  23. Gama J, Medas P, G Castillo, Rodrigues P (2004) Learning with drift detection. Advances in artificial intelligence—SBIA 2004. Springer, New York, pp 286–295CrossRefGoogle Scholar
  24. Gomes JB, Menasalvas E, Sousa PAC (2011) Learning recurring concepts from data streams with a context-aware ensemble. In: Proceedings of the 2011 ACM symposium on applied computing, SAC ’11, ACM, New York, pp 994–999. doi: 10.1145/1982185.1982403
  25. Hoens TR, Chawla NV, Polikar R (2011) Heuristic updatable weighted random subspaces for non-stationary environments. In Diane JC, Jian P, Wei W, Osmar RZ, Xindong W (ed), IEEE international conference on data mining, ICDM-11, IEEE, pp 241–250Google Scholar
  26. Hoens TR, Polikar R, Chawla NV (2012) Learning from streaming data with concept drift and imbalance: an overview. Prog Artif Intell 1(1):89–101. doi: 10.1007/s13748-011-0008-0 CrossRefGoogle Scholar
  27. Huang DTJ, Koh YS, Gillian D, Pears R (2013) Tracking drift types in changing data streams. In: Hiroshi M, Wu Z, Cao L, Zaiane O, Min Y, Wei W (eds) Advanced data mining and applications. Lecture notes in computer science. Springer, Berlin, pp 72–83. doi: 10.1007/978-3-642-53914-5_7
  28. Hulten G, Spencer L, Domingos P (2001) Mining time-changing data streams. In: Proceedings of the seventh ACM SIGKDD international conference on knowledge discovery and data mining, KDD-01, ACM, pp 97–106Google Scholar
  29. Jiang N, Gruenwald L (2006) Research issues in data stream association rule mining. ACM SIGMOD Rec 35(1):14–19CrossRefGoogle Scholar
  30. Kelly MG, Hand DJ, Adams NM (1999) The impact of changing populations on classifier performance. In: Proceedings of the fifth ACM SIGKDD international conference on knowledge discovery and data mining, KDD-99, New York, ACM, pp 367–371. doi: 10.1145/312129.312285
  31. Kosina Petr, Gama João, Sebastião Raquel (2010) Drift severity metric. European Conference on Artificial Intelligence, ECAI 2010:1119–1120Google Scholar
  32. Krempl G, Zliobaite I, Brzezinski D, Hullermeier E, Last M, Lemaire V, Noack T, Shaker A, Sievi S, Spiliopoulou M, Stefanowski J (2014) Open challenges for data stream mining research. In: ACM SIGKDD explorations newsletter, vol 16–1, pp 1–10Google Scholar
  33. Kuh A, Petsche T, Rivest RL (1991) Learning time-varying concepts. In: Advances in neural information processing systems, pp 183–189Google Scholar
  34. Kullback S, Leibler RA (1951) On information and sufficiency. Ann Math Stat 22(1):79–86Google Scholar
  35. Kuncheva LI (2004) Classifier ensembles for changing environments. In: Multiple Classifier Systems. Springer, pp 1–15Google Scholar
  36. Masud MM, Gao J, Khan L, Han J, Thuraisingham B (2011) Classification and novel class detection in concept-drifting data streams under time constraints. IEEE Trans Knowl Data Eng 23(6):859–874CrossRefGoogle Scholar
  37. Michalski RS (1983) A theory and methodology of inductive learning. Springer, New YorkGoogle Scholar
  38. Minku FL, Yao X (2009) Using diversity to handle concept drift in on-line learning. In: International joint conference on neural networks, IJCNN-09, IEEE, pp 2125–2132Google Scholar
  39. Minku LL, White AP, Xin Y (2010) The impact of diversity on online ensemble learning in the presence of concept drift. IEEE Trans Knowl Data Eng 22(5):730–742. doi: 10.1109/TKDE.2009.156 ISSN 1041–4347CrossRefGoogle Scholar
  40. Moreno-Torres Jose G, Raeder Troy, Alaiz-Rodrguez Rocio, Chawla Nitesh V, Herrera Francisco (2012) A unifying view on dataset shift in classification. Pattern Recognit 45(1):521–530 ISSN 0031-3203CrossRefGoogle Scholar
  41. Narasimhamurthy A, Kuncheva L (2007) A framework for generating data to simulate changing environments. In: Proceedings of the 25th IASTED international multi-conference: artificial intelligence and applications, ACTA Press, 549: p 389Google Scholar
  42. Nguyen H-L, Woon Y-K, Ng W-K, Wan L (2012) Heterogeneous ensemble for feature drifts in data streams. In: Advances in knowledge discovery and data mining. Springer, pp 1–12Google Scholar
  43. Nguyen H-L, Woon Y-K, Ng W-K (2014) A survey on data stream clustering and classification. Knowl Inf Syst pp 1–35Google Scholar
  44. Nishida Kyosuke, Yamauchi K (2007) Detecting concept drift using statistical testing. In: Discovery Science, Springer, pp 264–269Google Scholar
  45. Oza NC, Russell S (2001) Online bagging and boosting. In: Artificial Intelligence and Statistics 2001, Morgan Kaufmann pp 105–112Google Scholar
  46. Pfahringer B, Holmes G, Kirkby R (2007) New options for Hoeffding trees. In: Mehmet O, John T (eds) AI 2007: advances in artificial intelligence, 4830th edn., Lecture notes in computer scienceSpringer, New York, pp 90–99. doi: 10.1007/978-3-540-76928-6_11
  47. Quionero-Candela J, Sugiyama M, Schwaighofer A, Lawrence ND (2009) Dataset shift in machine learning. The MIT Press, CambridgeGoogle Scholar
  48. Shaker A, Hullermeier E (2015) Recovery analysis for adaptive learning from non-stationary data streams. In: Neurocomputing, ScienceDirect, pp 250–264Google Scholar
  49. Subramaniam S, Palpanas T, Papadopoulos D, Kalogeraki V, Gunopulos D (2006) Online outlier detection in sensor data using non-parametric models. In: Proceedings of the 32nd international conference on very large data bases, VLDB Endowment, pp 187–198Google Scholar
  50. Tsymbal A (2004) The problem of concept drift: definitions and related work. Technical Report TCD-CS-2004-15, The University of Dublin, Trinity College, Department of Computer Science, DublinGoogle Scholar
  51. Wetzel L (2009) Types and tokens. In: Zalta EN (ed) The Stanford Encyclopedia of Philosophy.
  52. Wang H, Fan W, Yu PS, Han J (2003) Mining concept-drifting data streams using ensemble classifiers. In: Proceedings of the ninth ACM SIGKDD international conference on knowledge discovery and data mining, KDD-03, New York, ACM, pp 226–235. doi: 10.1145/956750.956778
  53. Wang H, Fan W, Yu PS, Han J (2003b) Mining concept-drifting data streams using ensemble classifiers. In: Proceedings of the ninth ACM SIGKDD international conference on knowledge discovery and data mining, KDD-03, ACM, pp 226–235Google Scholar
  54. Wang S, Minku LL, Ghezzi D, Caltabiano D, Tino P, Yao X (2013) Concept drift detection for online class imbalance learning. In: The 2013 international joint conference on neural Network, IJCNN-13, IEEE, pp 1–10Google Scholar
  55. Widmer G, Kubat M (1996) Learning in the presence of concept drift and hidden contexts. Mach Learn 23(1):69–101. doi: 10.1007/BF00116900 ISSN 0885–6125Google Scholar
  56. Zhang P, Zhu X, Shi Y (2008) Categorizing and mining concept drifting data streams. In: Proceeding of the 14th ACM SIGKDD international conference on knowledge discovery and data mining, KDD-08, ACM, pp 812–820. doi: 10.1145/1401890.1401987
  57. Zliobaite I (2010) Learning under concept drift: an overview. Technical reportGoogle Scholar
  58. Zliobaite I (2014) Controlled permutation for testing adaptive learning models. Knowledge and information systems, vol 39. Springer, London, pp 565–578Google Scholar

Copyright information

© The Author(s) 2016

Authors and Affiliations

  1. 1.Faculty of Information TechnologyMonash UniversityClaytonAustralia
  2. 2.McLaren Applied Technologies Pte Ltd APACSingaporeSingapore

Personalised recommendations