Advertisement

Inaccuracies of Shape Averaging Method Using Dynamic Time Warping for Time Series Data

  • Vit Niennattrakul
  • Chotirat Ann Ratanamahatana
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4487)

Abstract

Shape averaging or signal averaging of time series data is one of the prevalent subroutines in data mining tasks, where Dynamic Time Warping distance measure (DTW) is known to work exceptionally well with these time series data, and has long been demonstrated in various data mining tasks involving shape similarity among various domains. Therefore, DTW has been used to find the average shape of two time series according to the optimal mapping between them. Several methods have been proposed, some of which require the number of time series being averaged to be a power of two. In this work, we will demonstrate that these proposed methods cannot produce the real average of the time series. We conclude with a suggestion of a method to potentially find the shape-based time series average.

Keywords

Time Series Shape Averaging Dynamic Time Warping 

References

  1. 1.
    Abdulla, W.H., Chow, D., Sin, G.: Cross-words reference template for DTW-based speech recognition systems. In: Proc. of TENCON 2003 (2003)Google Scholar
  2. 2.
    Bagnall, A., Janacek, G.: Clustering Time Series with Clipped Data. Mach. Learn. 58, 151–178 (2005)zbMATHCrossRefGoogle Scholar
  3. 3.
    Boudaoud, S., Rix, H., Meste, O.: Integral shape averaging and structural average estimation. IEEE Trans. on Signal Processing 53(10), 3644–3650 (2005)CrossRefMathSciNetGoogle Scholar
  4. 4.
    Bradley, P.S., Fayyad, U.M.: Refining Initial Points for K–Means Clustering. In: Proceedings of the 15th Int’l Conference on Machine Learning, pp. 91–99 (1998)Google Scholar
  5. 5.
    Caiani, E.G., Porta, A., Baselli, G., Turiel, M., Muzzupappa, S., Pieruzzi, F., Crema, C., Malliani, A., Cerutti, S.: Warped-average template technique to track on a cycle-bycycle basis the cardiac filling phases on left ventricular volume. IEEE Computers in Cardiology (1998)Google Scholar
  6. 6.
    Corradini, A.: Dynamic Time Warping for Off-Line Recognition of a Small Gesture Vocabulary. In: Proc. of the IEEE ICCV Workshop on Ratfg-Rts, Washington, DC (2001)Google Scholar
  7. 7.
    Chu, S., Keogh, E., Hart, D., Pazzani, M.: Iterative deepening dynamic time warping for time series. In: Proceedings of SIAM International Conference on Data Mining (2002)Google Scholar
  8. 8.
    Gupta, L., Molfese, D.L., Tammana, R., Simos, P.G.: Nonlinear alignment and averaging for estimating the evoked potential. IEEE Trans. on Biomed. Eng. 43(4), 348–356 (1996)CrossRefGoogle Scholar
  9. 9.
    Hu, J., Ray, B.: An Interleaved HMM/DTW Approach to Robust Time Series Clustering. IBM T.J. Watson Research Center (2006)Google Scholar
  10. 10.
    Keogh, E., Smyth, P.: An enhanced representation of time series which allows fast classification, clustering and relevance feedback. In: KDD, pp. 24–30 (1997)Google Scholar
  11. 11.
    Lange, D.H., Pratt, H., Inbar, G.F.: Modeling and estimation of single evoked brain potential components. IEEE Trans. on Biomed. Eng. 44, 791–799 (1997)CrossRefGoogle Scholar
  12. 12.
    Mor-Avi, V., Gillesberg, I.E., Korcarz, C., Sandelski, J., Lang, R.M.: Signal averaging helps reliable noninvasive monitoring of left ventricular dimensions based on acoustic quantification. Computers in Cardiology, 21–24 (1994)Google Scholar
  13. 13.
    Oates, T., Firoiu, L., Cohen, P.R.: Using Dynamic Time Warping to Bootstrap HMM-Based Clustering of Time Series. In: Sun, R., Giles, C.L. (eds.) IJCAI-WS 1999. LNCS (LNAI), vol. 1828, pp. 35–52. Springer, Heidelberg (2001)CrossRefGoogle Scholar
  14. 14.
    Rabiner, L.R., Levinson, S.E., Rosenberg, A.E., Wilpon, J.G.: Speaker-independent recognition of isolated words using clustering techniques. In: Readings in Speech Recognition, CA, 166–179 (1990)Google Scholar
  15. 15.
    Ratanamahatana, C.A., Keogh, E.: Everything you know about Dynamic Time Warping is Wrong. In: Proc. of KDD Workshop on Mining Temporal and Sequential Data (2004)Google Scholar
  16. 16.
    Ratanamahatana, C.A., Keogh, E.J.: Multimedia Retrieval Using Time Series Representation and Relevance Feedback. In: Fox, E.A., Neuhold, E.J., Premsmit, P., Wuwongse, V. (eds.) ICADL 2005. LNCS, vol. 3815, pp. 400–405. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  17. 17.
    Saito, N.: Local feature extraction and its application using a library of bases. PhD thesis, Yale University (1994)Google Scholar
  18. 18.
    Salvador, S., Chan, P.: FastDTW: Toward Accurate Dynamic Time Warping in Linear Time and Space. In: Proc. of KDD Workshop on Mining Temporal and Sequential Data (2004)Google Scholar
  19. 19.
    Wilpon, J., Rabiner, L.: A modified K-means clustering algorithm for use in isolated work recognition. IEEE Trans. on Signal Processing. 33, 587–594 (1985)CrossRefGoogle Scholar

Copyright information

© Springer Berlin Heidelberg 2007

Authors and Affiliations

  • Vit Niennattrakul
    • 1
  • Chotirat Ann Ratanamahatana
    • 1
  1. 1.Department of Computer Engineering, Chulalongkorn University, Phayathai Rd., Pathumwan, Bangkok 10330Thailand

Personalised recommendations