Unsupervised outlier detection for time series by entropy and dynamic time warping

  • Seif-Eddine Benkabou
  • Khalid Benabdeslem
  • Bruno Canitia
Regular Paper
  • 126 Downloads

Abstract

In the last decade, outlier detection for temporal data has received much attention from data mining and machine learning communities. While other works have addressed this problem by two-way approaches (similarity and clustering), we propose in this paper an embedded technique dealing with both methods simultaneously. We reformulate the task of outlier detection as a weighted clustering problem based on entropy and dynamic time warping for time series. The outliers are then detected by an optimization problem of a new proposed cost function adapted to this kind of data. Finally, we provide some experimental results for validating our proposal and comparing it with other methods of detection.

Keywords

Anomaly detection Time series DTW Weighted clustering 

Notes

Acknowledgements

We thank anonymous reviewers for their very useful comments and suggestions.

Compliance with ethical standards

Conflicts of interest

The authors declare that they have no conflict of interest.

References

  1. 1.
    Aggarwal C (2013) Outlier analysis. Springer, BerlinCrossRefMATHGoogle Scholar
  2. 2.
    Aggarwal C, Zhao Y, Yu P (2011) Outlier detection in graph streams. In: Proceedings of ICDE, pp 399–409Google Scholar
  3. 3.
    Aggarwal C, Subbian K (2012) Event detection in social streams. In: Proceedings of SDM, pp 624–635Google Scholar
  4. 4.
    Bahadori M, Kale D, Yingying F, Yan L (2015) Functional subspace clustering with application to time series. Proceedings of ICML, pp. 228–237Google Scholar
  5. 5.
    Basu S, Meckesheimer M (2007) Automatic outlier detection for time series: an application to sensor data. Knowl Inf Syst 11(2):137–154CrossRefGoogle Scholar
  6. 6.
    Bevilacqua M, Tsaftaris S (2015) Dictionary-decomposition-based one-class svm for unsupervised detection of anomalous time series. In: Proceedings of 23rd European signal processing conference (EUSIPCO), pp 1776–1780Google Scholar
  7. 7.
    Budalakoti S, Srivastava A, Otey M (2009) Anomaly detection and diagnosis algorithms for discrete symbol sequences with applications to airline safety. IEEE Trans Syst Man Cybern Part C Appl 39(1):101–113CrossRefGoogle Scholar
  8. 8.
    Chandola V, Mithal V, Kumar V (2008) Comparative evaluation of anomaly detection techniques for sequence data. In: Proceedings of ICDM, pp 743–748Google Scholar
  9. 9.
    Chen Y, Keogh E, Hu B, Begum N, Bagnall A, Mueen A, Batista G (2015) The ucr time series classification archive. www.cs.ucr.edu/~eamonn/time_series_data/
  10. 10.
    Dasgupta D, Nino F (2000) A comparison of negative and positive selection algorithms in novel pattern detection. In: Proceedings of IEEE International Conference on Systems, Man, and Cybernetics, pp 125–130Google Scholar
  11. 11.
    Demšar J (2006) Statistical comparisons of classifiers over multiple data sets. J Mach Learn Res 7:1–30MathSciNetMATHGoogle Scholar
  12. 12.
    Dit-Yan Y, Calvin C (2002) Parzen-window network intrusion detectors. In: Proceedings of ICPR, pp 385–388Google Scholar
  13. 13.
    Fawcett T (2006) An introduction to ROC analysis. Pattern Recognit Lett 27(8):861–874MathSciNetCrossRefGoogle Scholar
  14. 14.
    Ferraty F, Vieu P (2006) Nonparametric functional data analysis: theory and practice. Springer, BerlinMATHGoogle Scholar
  15. 15.
    Fox A (1972) Outliers in time series. J R Stat Soc Ser B Methodol 34(3):350–363MathSciNetMATHGoogle Scholar
  16. 16.
    Gao B, Ma H, Yang Y (2002) Hmms (hidden markov models) based on anomaly intrusion detection method. In: Proceedings of Conference on machine learning and cybernetics, pp 381–385Google Scholar
  17. 17.
    Gao J, Liang F, Fan W, Wang C, Sun Y, Han J (2010) On community outliers and their efficient detection in information networks. In: Proceedings of KDD, pp 813–822Google Scholar
  18. 18.
    Goldstein M, Uchida S (2016) A comparative evaluation of unsupervised anomaly detection algorithms for multivariate data. PLoS ONE 11(4):1–31Google Scholar
  19. 19.
    Görnitz N, Braun L, Kloft M (2015) Hidden markov anomaly detection. In: Proceedings of ICML, pp 1833–1842Google Scholar
  20. 20.
    Green P, Kim J, Carmone F (1990) A preliminary study of optimal variable weighting in k-means clustering. J Classif 7(2):271–285CrossRefGoogle Scholar
  21. 21.
    Gupta M, Gao J, Aggarwal C, Han J (2014) Outlier detection for temporal data: a survey. IEEE Trans Knowl Data Eng 26(9):2250–2267CrossRefMATHGoogle Scholar
  22. 22.
    Gupta M, Gao J, Sun Y, Han J (2012) Integrating community matching and outlier detection for mining evolutionary community outliers. In: Proceedings of KDD, pp 859–867Google Scholar
  23. 23.
    Gupta M, Gao J, Sun Y, Han J (2012) Community trend outlier detection using soft temporal pattern mining. In: Proceedings of ECML/PKDD, pp 692–708Google Scholar
  24. 24.
    Hautamaki T, Nykanen P, Frant P (2008) Time-series clustering by approximate prototypes. In: Proceedings of ICPR, pp 1–4Google Scholar
  25. 25.
    Hawkins D (1980) Identification of outliers. Chapman and Hall, LondonCrossRefMATHGoogle Scholar
  26. 26.
    Huang J, Ng M, Rong H, Li Z (2005) Automated variable weighting in k-means type clustering. IEEE Trans Pattern Anal Mach Intell 27:657–668CrossRefGoogle Scholar
  27. 27.
    Jing L, Ng M, Huang Z (2007) An entropy weighting k-means algorithm for subspace clustering of high-dimensional sparse data. IEEE Trans Knowl Data Eng 19(8):1026–1041CrossRefGoogle Scholar
  28. 28.
    Kassab R, Alexandre F (2009) Incremental data-driven learning of a novelty detection model for one-class classification with application to high-dimensional noisy data. Mach Learn 74(2):191–234CrossRefMATHGoogle Scholar
  29. 29.
    Keogh E, Lin J, Lee S, Herle V (2006) Finding the most unusual time series subsequence: algorithms and applications. Knowl Inf Syst 11(1):1–27CrossRefGoogle Scholar
  30. 30.
    Lane T, Brodley C (1997) Sequence matching and learning in anomaly detection for computer security. AI Approaches to Fraud Detection and Risk Management. In: AAAI Workshop, pp 43–49Google Scholar
  31. 31.
    Lee Y, Yeh R, Wang F (2013) Anomaly detection via online oversampling principal component analysis. IEEE Trans Knowl Data Eng 25(7):1460–1470CrossRefGoogle Scholar
  32. 32.
    Makarenkov V, Legendre P (2001) Optimal variable weighting for ultrametric and additive trees and k-means partitioning: methods and software. J Classif 18:245–271MathSciNetMATHGoogle Scholar
  33. 33.
    Markus M, Hans-Peter K, Raymond T, Jrg S (2000) LOF: identifying density-based local outliers. In: Proceedings of SIGMOD Conference, pp 93–104Google Scholar
  34. 34.
    Modha D, Spangler S (2003) Feature weighting in k-means clustering. Mach Learn 52:217–237CrossRefMATHGoogle Scholar
  35. 35.
    Ng A, Jordan M, Weiss Y (2002) On spectral clustering: analysis and an algorithm. In: Proceedings of neural information processing systems (NIPS), pp 849–856. MIT PressGoogle Scholar
  36. 36.
    Palpanas T, Papadopoulos D, Kalogeraki V, Gunopulos D (2003) Distributed deviation detection in sensor networks. Proc SIGMOD Rec 32(4):77–82CrossRefGoogle Scholar
  37. 37.
    Paulheim H, Meusel R (2015) A decomposition of the outlier detection problem into a set of supervised learning problems. Mach Learn 100(2–3):509–531MathSciNetCrossRefMATHGoogle Scholar
  38. 38.
    Petitjean F, Forestier G, Webb G, Nicholson A, Chen Y, Keogh E (2014) Dynamic time warping averaging of time series allows faster and more accurate classification. Proc ICDM 2014(2014):470–479Google Scholar
  39. 39.
    Portnoy L, Eskin E, Stolfo S (2001) Intrusion detection with unlabeled data using clustering. In: Proceedings of ACM CSS Workshop on Data Mining Applied to Security (DMSA), pp 5–8Google Scholar
  40. 40.
    Rakthanmanon T, Campana B, Mueen A, Batista G, Westover M, Zhu Q, Zakaria J, Keogh E (2012) Searching and mining trillions of time series subsequences under dynamic time warping. In: Proceedings of ACM SIGKDD, pp 262–270Google Scholar
  41. 41.
    Ratanamahatana C, Keogh E (2004) Making time-series classification more accurate using learned constraints. Proc SIAM 2004:11–22MathSciNetGoogle Scholar
  42. 42.
    Rebbapragada U, Protopapas P, Brodley C, Alcock C (2009) Finding anomalous periodic time series. Mach Learn 74(3):281–313CrossRefGoogle Scholar
  43. 43.
    Salvador S, Chan P (2004) Fastdtw: Toward accurate dynamic time warping in linear time and space. In: Proceedings of KDD workshop on mining temporal and sequential data, pp 70–80Google Scholar
  44. 44.
    Salvador S, Chan P (2005) Learning states and rules for detecting anomalies in time series. Appl Intell 23(3):241–255CrossRefGoogle Scholar
  45. 45.
    Salvador S, Chan P (2007) Toward accurate dynamic time warping in linear time and space. Intell Data Anal 11(5):561–580Google Scholar
  46. 46.
    Schölkopf B, Williamson R, Smola A, Shawe-Taylor J, Platt J (1999) Support vector method for novelty detection. In: Proceedings of neural information processing systems (NIPS), pp 582–588Google Scholar
  47. 47.
    Shang H (2014) A survey of functional principal component analysis. Adv Stat Anal 98(2):121–142MathSciNetCrossRefGoogle Scholar
  48. 48.
    Tian S, Mu S, Yin C (2007) Sequence-similarity kernels for svms to detect anomalies in system calls. Neurocomputing 70(4–6):859–866CrossRefGoogle Scholar
  49. 49.
    Vintsyuk T (1968) Speech discrimination by dynamic programming. Cybernetics 4(1):52–57MathSciNetCrossRefGoogle Scholar

Copyright information

© Springer-Verlag London 2017

Authors and Affiliations

  • Seif-Eddine Benkabou
    • 1
    • 2
  • Khalid Benabdeslem
    • 1
  • Bruno Canitia
    • 2
  1. 1.University of Lyon1-LIRISVilleurbanneFrance
  2. 2.LIZEO Online Media GroupLyonFrance

Personalised recommendations