Clustering Time Series with Granular Dynamic Time Warping Method

In this paper, a new method, named granular dynamic time warping is proposed. This method is based on the granular approach of information granulation and has the characteristics of dynamic time warping approach. Thus it can be used to cluster time series with different lengths on the granular level. To cluster time series, this method first builds the corresponding granular time series, and then does the clustering on the granular time series. With this method, higher efficiency will be achieved in clustering time series, which is a goal pursued in clustering of large amount of time series. We also illustrate the prior performance of the new method with experiments.

[1]  Zhu Zhong-ying.  Novel Algorithm for Time Series Data Mining Based on Dynamic Time Warping , 2004 .

[2]  S. Levinson,et al.  Considerations in dynamic time warping algorithms for discrete word recognition , 1978 .

[3]  Li Yuan.  A Study on Dynamic Time Warping , 2002 .

[4]  Fusheng Yu,et al.  A granulation-based method for finding similarity between time series , 2005, 2005 IEEE International Conference on Granular Computing.

[5]  G. W. Hughes,et al.  Minimum Prediction Residual Principle Applied to Speech Recognition , 1975 .

[6]  Li Yuan.  A Study on Dynamic Time Warping Based on Different Constraints , 2002 .

[7]  C. Myers,et al.  A level building dynamic time warping algorithm for connected word recognition , 1981 .

[8]  Witold Pedrycz,et al.  Granular computing: an introduction , 2001, Proceedings Joint 9th IFSA World Congress and 20th NAFIPS International Conference (Cat. No. 01TH8569).

[9]  Lotfi A. Zadeh,et al.  Fuzzy sets and information granularity , 1996 .

[10]  Eamonn J. Keogh,et al.  UCR Time Series Data Mining Archive , 1983 .

[11]  W. Pedrycz,et al.  Granulation of temporal data: a global view on time series , 2003, 22nd International Conference of the North American Fuzzy Information Processing Society, NAFIPS 2003.

[12]  Witold Pedrycz,et al.  The design of fuzzy information granules: Tradeoffs between specificity and experimental evidence , 2009, Appl. Soft Comput..

[13]  Heikki Mannila,et al.  Rule Discovery from Time Series , 1998, KDD.

[14]  David Malah,et al.  Dynamic time warping with path control and non-local cost , 1994, Proceedings of the 12th IAPR International Conference on Pattern Recognition, Vol. 2 - Conference B: Computer Vision & Image Processing. (Cat. No.94CH3440-5).