Data Mining and Knowledge Discovery

, Volume 14, Issue 1, pp 99–129 | Cite as

Compression-based data mining of sequential data

  • Eamonn KeoghEmail author
  • Stefano Lonardi
  • Chotirat Ann Ratanamahatana
  • Li Wei
  • Sang-Hee Lee
  • John Handley


The vast majority of data mining algorithms require the setting of many input parameters. The dangers of working with parameter-laden algorithms are twofold. First, incorrect settings may cause an algorithm to fail in finding the true patterns. Second, a perhaps more insidious problem is that the algorithm may report spurious patterns that do not really exist, or greatly overestimate the significance of the reported patterns. This is especially likely when the user fails to understand the role of parameters in the data mining process. Data mining algorithms should have as few parameters as possible. A parameter-light algorithm would limit our ability to impose our prejudices, expectations, and presumptions on the problem at hand, and would let the data itself speak to us. In this work, we show that recent results in bioinformatics, learning, and computational theory hold great promise for a parameter-light data-mining paradigm. The results are strongly connected to Kolmogorov complexity theory. However, as a practical matter, they can be implemented using any off-the-shelf compression algorithm with the addition of just a dozen lines of code. We will show that this approach is competitive or superior to many of the state-of-the-art approaches in anomaly/interestingness detection, classification, and clustering with empirical tests on time series/DNA/text/XML/video datasets. As a further evidence of the advantages of our method, we will demonstrate its effectiveness to solve a real world classification problem in recommending printing services and products.


Kolmogorov complexity Parameter-free data mining Anomaly detection Clustering 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Allison L, Stern L, Edgoose T, Dix TI (2000) Sequence complexity for biological sequence analysis. Comput Chem 24(1):43–55Google Scholar
  2. Baronchelli A, Caglioti E, Loreto V (2005) Artificial sequences and complexity measures. J. Stat. Mech: Theory and Exp, Issue 04, P04002Google Scholar
  3. Benedetto D, Caglioti E, Loreto V (2002) Language trees and zipping. Phys Rev Lett 88: 048702CrossRefGoogle Scholar
  4. Chakrabarti D, Papadimitriou S, Modha D, Faloutsos C (2004) Fully automatic cross-assocations, In: Proceedings of the KDD 2004, Seattle, WAGoogle Scholar
  5. Christen P, Goiser K (2005) Towards automated data linkage and deduplication. Tech Report, Australian National UniversityGoogle Scholar
  6. Cook D, Holder LB (2000) Graph-based data mining. IEEE Intell Syst 15(2):32–41CrossRefGoogle Scholar
  7. Dasgupta D, Forrest S (1999) Novelty detection in time series data using ideas from immunology. In: Proc. of the international conference on intelligent systems, Heidelberg, GermanyGoogle Scholar
  8. Domingos P (1998) A process-oriented heuristic for model selection. In: Machine learning Proc. of the fifteenth international conference,. Morgan Kaufmann Publishers, San Francisco, CA, pp 27–135Google Scholar
  9. Elkan, C (2001) Magical thinking in data mining: lessons from CoIL challenge 2000. In Proc. of SIGKDD 2001, San Francisco, CA, USA, pp 426–431Google Scholar
  10. Elkan C (2003) Using the triangle inequality to accelerate k-Means. In: Proc. of ICML 2003, Washington DC, USA, pp 147–153Google Scholar
  11. Faloutsos C, Lin K (1995) FastMap: a fast algorithm for indexing, data-mining and visualization of traditional and multimedia datasets. In Proc. of 24th ACM SIGMOD, San Jose, CA, USAGoogle Scholar
  12. Farach M, Noordewier M, Savari S, Shepp L, Wyner A, Ziv J (1995) On the entropy of DNA: algorithms and measurements based on memory and rapid convergence. In: Proc. of the symp. on discrete algorithms, San Francisco, CA, USA pp 48-57Google Scholar
  13. Ferrandina F, Meyer T, Zicari R (1994) Implementing lazy database updates for an object database system. In: Proc. of the 20 international conference on very large databases, Santiago de Chile, Chile, pp 261–272Google Scholar
  14. Flexer A (1996) Statistical evaluation of neural networks experiments: minimum requirements and current practice. In: Proc. of the 13th european meeting on cybernetics and systems research, vol. 2, Austria, pp 1005–1008Google Scholar
  15. Frank E, Chui C, Witten I (2000) Text categorization using compression models. In: Proc. of the IEEE data compression conference, Snowbird, Utah, IEEE Comput Soc p555Google Scholar
  16. Gaussier E, Goutte C, Popat K, Chen F (2002) A hierarchical model for clustering and categorising documents source lecture notes in computer science; Vol. 2291 archive Proceedings of the 24th BCS-IRSG european colloquium on IR research: advances in information retrieval, Glasgow, UKGoogle Scholar
  17. Gatlin L (1972) Information theory and the living systems. Columbia University Press, columbiaGoogle Scholar
  18. Gavrilov M, Anguelov D, Indyk P, Motwahl R (2000) Mining the stock market: which measure is best? In: Proc. of the 6th ACM SIGKDD, 2000, Boston, MA, USAGoogle Scholar
  19. Ge X, Smyth P (2000) Deformable Markov model templates for time-series pattern matching. In: Proc. of the 6th ACM SIGKDD, Boston, MA, pp 81–90Google Scholar
  20. Goldberger A.L, Amaral L, Glass L, Hausdorff JM, Ivanov PCh, Mark RG, Mietus JE, Moody GB, Peng CK, Stanley HE (2000) PhysioBank, physioToolkit, and physioNet: components of a new research resource for complex physiologic signals. Circulation 101(23):e215–e220Google Scholar
  21. Kalpakis K, Gada D, Puttagunta V (2001) Distance measures for effective clustering of ARIMA time-series. In: Proceedings of the 1st IEEE ICDM, San Jose, CA, pp 273-280Google Scholar
  22. Kennel M (2004) Testing time symmetry in time series using data compression dictionaries. Phys Rev E 69; 056208Google Scholar
  23. Keogh E.∼eamonn/SIGKDD2004, University of California, RiversideGoogle Scholar
  24. Keogh E, Folias T (2002) The UCR time series data mining archive. University of California, Riverside CA [∼eamonn/TSDMA/index.html]Google Scholar
  25. Keogh E, Kasetty S (2002) On the need for time series data mining benchmarks: a survey and empirical demonstration. In: Proc. of SIGKDD, Edmonton, Alberta, CanadaGoogle Scholar
  26. Keogh E, Lin J, Truppel W (2003) Clustering of time series subsequences is meaningless: implications for past and future research. In: Proc. of the 3rd IEEE ICDM, Melbourne, FL, pp 115–122Google Scholar
  27. Kit C (1998) A goodness measure for phrase learning via compression with the MDL principle. In: Kruijff-Korbayova I (ed) The ELLSSI-98 student session, Chapt 13, Saarbrueken, pp 175–187Google Scholar
  28. Li M, Badger JH, Chen X, Kwong S, Kearney P, Zhang H (2001) An information-based sequence distance and its application to whole mitochondrial genome phylogeny. Bioinformatics 17:149–154CrossRefGoogle Scholar
  29. Li M, Chen X, Li X, Ma B, Vitanyi, P (2003) The similarity metric. In: Proc. of the fourteenth annual ACM-SIAM symposium on Discrete algorithms, Baltimore, MD, USA, pp 863–872Google Scholar
  30. Li M, Vitanyi P (1997) An introduction to kolmogorov complexity and its applications, 2nd edn, Springer Verlag, BerlinGoogle Scholar
  31. Lin J, Keogh E, Lonardi S, Chiu B (2003) A symbolic representation of time series, with implications for streaming algorithms. In: Proc. of the 8th ACM SIGMOD workshop on research issues in data mining and knowledge discovery, San Diego, CAGoogle Scholar
  32. Loewenstern D, Hirsh H, Yianilos P, Noordewier M (1995) DNA sequence classification using compression-based induction, DIMACS Technical Report 95-04Google Scholar
  33. Loewenstern D, Yianilos PN (1999) Significantly lower entropy estimates for natural DNA sequences, J Comput Biol 6(1)Google Scholar
  34. Ma J, Perkins S (2003) Online novelty detection on temporal sequences. In: Proc. international conference on knowledge discovery and data mining, Washington, DCGoogle Scholar
  35. Mahoney M, Chan P (2005) Learning rules for time series anomaly detection. SensorMiner Tech report (available at [])Google Scholar
  36. Mehta M, Rissanen J, Agrawal R (1995) MDL-based decision tree pruning, In: Proceedings of the first international conference on knowledge discovery and data mining (KDD’95), Montreal, CanadaGoogle Scholar
  37. Needham S, Dowe D(2001) Message length as an effective ockham’s razor in decision tree induction, In: Proc. 8th international workshop on AI and statistics, Key West, FL, USA, pp 253–260Google Scholar
  38. Ortega A, Beferull-Lozano B, Srinivasamurthy N, Xie H (2000) Compression for recognition and content based retrieval. In: Proc. of the European signal processing conference, EUSIPCO’00, Tampere, FinlandGoogle Scholar
  39. Papadimitriou S, Gionis A, Tsaparas P, Väisänen A, Mannila H, Faloutsos C (2005) Parameter-free spatial data mining using MDL, In: Proc of the 5th International Conference on Data Mining (ICDM), Houston, TX, USAGoogle Scholar
  40. Quinlan JR, Rivest RL (1989) Inferring decision trees using the minimum description length principle. Infor Comput 80:227–248zbMATHCrossRefMathSciNetGoogle Scholar
  41. Ratanamahatana CA, Keogh E (2004) Making time-series classification more accurate using learned constraints. In: Proc. of SIAM international conference on data mining (SDM ’04), Lake Buena Vista, FloridaGoogle Scholar
  42. Rissanen J (1978) Modeling by shortest data description. Automatica, 14:465–471zbMATHCrossRefGoogle Scholar
  43. Salzberg SL (1997) On comparing classifiers: Pitfalls to avoid and a recommended approach. Data Min Knowl Disc 1(3):317–328CrossRefGoogle Scholar
  44. Segen J (1990) Graph clustering and model learning by data compression. In: Proc. of the machine learning conference, Austin, TX, USA, pp 93–101Google Scholar
  45. Sculley D, Brodley CE (2006) Compression and machine learning: a new perspective on feature space vectors, In: Proceedings of data compression conference, Snowbird, UT, USA, pp 332–341Google Scholar
  46. Shahabi C, Tian X, Zhao W (2000) TSA-tree: a wavelet-based approach to improve the efficiency of multi-level surprise and trend queries. In: Proc. of the 12th Int’l conference on scientific and statistical database management (SSDBM 2000), Berlin, GermanyGoogle Scholar
  47. Teahan WJ, Wen Y, McNab RJ, Witten IH (2000) A compression-based algorithm for Chinese word segmentation. Comput Linguist 26:375–393CrossRefGoogle Scholar
  48. Vlachos M, Hadjieleftheriou M, Gunopulos D, Keogh E (2003) Indexing multi-dimensional time-series with support for multiple distance measures. In: Proc. of the 9th ACM SIGKDD, Washington, DC, USA, pp 216–225Google Scholar
  49. Wallace C, Boulton (1968) An information measure for classification. Comput J 11 (2):185–194zbMATHGoogle Scholar
  50. Yairi T, Kato Y, Hori K (2001) Fault detection by mining association rules from house-keeping data. In: Proc. of Int’l sym. on AI, Robotics and Automation in SpaceGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2007

Authors and Affiliations

  • Eamonn Keogh
    • 1
    Email author
  • Stefano Lonardi
    • 1
  • Chotirat Ann Ratanamahatana
    • 2
  • Li Wei
    • 1
  • Sang-Hee Lee
    • 3
  • John Handley
    • 4
  1. 1.Department of Computer Science and EngineeringUniversity of CaliforniaRiversideUSA
  2. 2.Department of Computer EngineeringChulalongkorn UniversityBangkokThailand
  3. 3.Department of AnthropologyUniversity of CaliforniaRiversideUSA
  4. 4.Xerox Innovation Group, Xerox CorporationNew YorkUSA

Personalised recommendations