Abstract
Distance measures play a central role for time series data. Such measures condense two complex structures into a convenient, single number – at the cost of loosing many details. This might become a problem when the series are in general quite similar to each other and series from different classes differ only in details. This work aims at supporting an analyst in the explorative data understanding phase, where she wants to get an impression of how time series from different classes compare. Based on the interval tree of scales, we develop a visualisation that draws the attention of the analyst immediately to those details of a time series that are representative or discriminative for the class. The visualisation adopts to the human perception of a time series by adressing the persistence and distinctiveness of landmarks in the series.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
In order to get meaningful results with time warping methods, the value range of both series should clearly overlap, as it may be obtained from standardization (to zero mean and unit variance).
- 2.
As before, we weigh the cases such that the total weight of series from the same class and series from a different class becomes identical to get accuracies independent of the number of classes.
References
Aigner, W., Miksch, S., Schumann, H., Tominski, C.: Visualization of Time-Oriented Data. HCI. Springer, London (2011)
Alencar, A.B., Paulovich, F.V., Minghim, R., Filho, M.G.D.A., Oliveira, M.C.F.D.: Similarity-based visualization of time series collections: an application to analysis of streamows. In: International Conference Information Visualisation, pp. 280–286, July 2008
Berndt, D.J., Clifford, J.: Finding patterns in time series: a dynamic programming approach. In: Advances in Knowledge Discovery and Data Mining, pp. 229–248. MIT Press (1996)
Mallat, S.G.: A Wavelet Tour of Signal Processing. Elsevier Ltd., Amsterdam (2001)
Perng, C.-S., Wang, H., Zhang, S.R., Parker, D.S.: Landmarks: a new model for similarity-based pattern querying in time series databases. In: International Conference on Data Engineering (2000)
Wang, X., Mueen, A., Ding, H., Trajcevski, G., Scheuermann, P., Keogh, E.: Experimental comparison of representation methods and distance measures for time series data. Data Min. Knowl. Disc. 26(2), 275–309 (2012)
Witkin, A.P.: Scale space filtering. In: Artifical Intelligence (IJCAI), Karlsruhe, Germany, pp. 1019–1022 (1983)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing AG
About this paper
Cite this paper
Sobek, T., Höppner, F. (2016). Visual Perception of Discriminative Landmarks in Classified Time Series. In: Boström, H., Knobbe, A., Soares, C., Papapetrou, P. (eds) Advances in Intelligent Data Analysis XV. IDA 2016. Lecture Notes in Computer Science(), vol 9897. Springer, Cham. https://doi.org/10.1007/978-3-319-46349-0_7
Download citation
DOI: https://doi.org/10.1007/978-3-319-46349-0_7
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-46348-3
Online ISBN: 978-3-319-46349-0
eBook Packages: Computer ScienceComputer Science (R0)