StreamKrimp: Detecting Change in Data Streams

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5211)


Data streams are ubiquitous. Examples range from sensor networks to financial transactions and website logs. In fact, even market basket data can be seen as a stream of sales. Detecting changes in the distribution a stream is sampled from is one of the most challenging problems in stream mining, as only limited storage can be used. In this paper we analyse this problem for streams of transaction data from an MDL perspective. Based on this analysis we introduce the StreamKrimp algorithm, whichuses the Krimp algorithm to characterise probability distributions with code tables. With these code tables, StreamKrimp partitions the stream into a sequence of substreams. Each switch of code table indicates a change in the underlying distribution. Experiments on both real and artificial streams show that StreamKrimp detects the changes while using only a very limited amount of data storage.


Data Stream Minimum Description Length Concept Drift Code Table Stream Mining 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Aggarwal, C.C.: A framework for diagnosing changes in evolving data streams. In: Proceedings of ACM SIGMOD 2003 (2003)Google Scholar
  2. 2.
    Aggarwal, C.C.: On Abnormality Detection in Spuriously Populated Data Streams. In: Proceedings of SIAM Conference on Data Mining 2005 (2005)Google Scholar
  3. 3.
    Aggarwal, C.C. (ed.): Data Streams: Models and Algorithms. Springer, Heidelberg (2007)zbMATHGoogle Scholar
  4. 4.
    Calders, T., Dexters, N., Goethals, G.: Mining Frequent Itemsets in a Stream. In: Proceedings of IEEE ICDM 2007 (2007)Google Scholar
  5. 5.
    Chen, K., Liu, L.: Detecting the Change of Clustering Structure in Categorical Data Streams. In: Proceedings of SIAM Conference on Data Mining 2006 (2006)Google Scholar
  6. 6.
    Coenen, F. The LUCS-KDD Discretised/normalised ARM and CARM Data Library (2003),
  7. 7.
    Dasu, T., Krishnan, S., Venkatasubramanian, S., Yi, K.: An Information-Theoretic Approach to Detecting Changes in Multi-Dimensional Data Streams. In: Proceedings of Interface 2006 (2006)Google Scholar
  8. 8.
    Geurts, K., Wets, G., Brijs, T., Vanhoof, K.: Profiling of high-frequency accident locations using association rules. In Transportation research record 1840 (2003)Google Scholar
  9. 9.
    Grünwald, P.D.: Minimum description length tutorial. In: Grünwald, P.D., Myung, I.J., Pitt, M.A. (eds.) Advances in Minimum Description Length. MIT Press, Cambridge (2005)Google Scholar
  10. 10.
    Kifer, D., Ben-David, S., Gehrke, J.: Detecting Change in Data Streams. In: Proceedings of VLDB 2004 (2004)Google Scholar
  11. 11.
    Muthukrishnan, S., van den Berg, E., Wu, Y.: Sequential Change Detection on Data Streams. In: Proceedings of the ICDM Workshops 2007 (2007)Google Scholar
  12. 12.
    Siebes, A., Vreeken, J., Van Leeuwen, M.: Item Sets That Compress. In: Proc. of the ACM SIAM Conference on Data Mining, pp. 393–404 (2006)Google Scholar
  13. 13.
    Papadimitriou, S., Brockwell, A., Faloutsos, C.: Adaptive, unsupervised stream mining. The VLDB Journal 13(3), 222–239 (2004)CrossRefGoogle Scholar
  14. 14.
    Vreeken, J., Van Leeuwen, M., Siebes, A.: Characterising the Difference. In: Proceedings of ACM SIGKDD 2007 (2007)Google Scholar
  15. 15.
    Vreeken, J., Van Leeuwen, M., Siebes, A.: Preserving Privacy through Generation. In: Proceedings of IEEE ICDM 2007 (2007)Google Scholar
  16. 16.
    Widmer, G., Kubat, M.: Learning in the Presence of Concept Drift and Hidden Contexts. Machine Learning 23, 69–101 (1996)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  1. 1.Department of Computer ScienceUniversiteit Utrecht 

Personalised recommendations