Querying and mining of time series data: experimental comparison of representations and distance measures

By Hui Ding, Goce Trajcevski, Peter Scheuermann, Xiaoyue Wang, Eamonn J. Keogh. Proceedings of the VLDB Endowment, 2008

1,489 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/8332d7dfa45120e880acf81694204ece421e8270

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Signal Processing

Citations in the corpus

Learning time-series shapelets (2014, methodology): “• Other Related Methods: The Fast Shapelets (FSH) [12] exploits a fast random projection technique on the SAX representation, while the Dynamic Time Warping (DTW) classifier on the raw time-series data is also selected due to its reputation as a strong similarity metric [4].”
Time series shapelets: a new primitive for data mining (2009, neutral): “Time Series Shapelets: A New Primitive for Data Mining Lexiang Ye Dept. of Computer Science & Engineering University of California, Riverside, CA 92521 lexiangy@cs.ucr.edu ABSTRACT Classification of time series has been attracting great interest over the past decade.”
Spatio-Temporal Data Mining (2017, background): “…been performed in limited contexts for specific topics, e.g., for trajectory pattern mining [Giannotti et al. 2007; Mamoulis et al. 2004] and time-series data mining [Ding et al. 2008; Esling and Agon 2012], an extensive evaluation covering the full breadth of STDM problems and methods is missing.”
A review on time series data mining (2011, background): “Finally, detail comparisons or experiments on the existing time series representation and similarity measure approaches can be found in Keogh and Kasetty (2002) and Ding et al. (2008) .”
Learning Maximum Margin Temporal Warping for Action Recognition (2013, methodology): “This cost function can be regarded as a soft-version of the commonly used Sakoe-Chiba Band constraint [4].”
Highly Comparative Feature-Based Time-Series Classification (2014, methodology): “using lower bounds [1], [5], [35], and speedups can be obtained using indexing [36], [37], [38] or time-domain dimensionality reduction [39], the need to calculate many distances between pairs of time series is fundamental to”
Using dynamic time warping distances as features for improved time series classification (2016, methodology): “…that have been proposed for time series classification, the decades old dynamic time warping (DTW) in combination with one-nearest neighbor has not been easy to beat convincingly (Xi et al. 2006; Ding et al. 2008; Xing et al. 2010; Chen et al. 2013; Wang et al. 2013; Lines and Bagnall 2014).”
Patterns of temporal variation in online media (2011, background): “Two key components of time series clustering are a distance measure [11], and a clustering algorithm [32].”
Exploiting multi-channels deep convolutional neural networks for multivariate time series classification (2015, methodology): “In the literature, DTW is considered as an alternative that is more robust than Euclidean distance and can align two sequences effectively [2, 9, 10].”
A Bag-of-Features Framework to Classify Time Series (2013, methodology): “We use the same approach proposed by [12] to compare results.”
Time Series Classification Using Multi-Channels Deep Convolutional Neural Networks (2014, methodology): “Among these classification methods, the distance-based method k-Nearest Neighbor (k-NN) classification has been empirically proven to be very difficult to beat [4, 5].”
Generalizing DTW to the multi-dimensional case requires an adaptive approach (2016, positive): “However, the DTW distance subsumes the Euclidean distance as a special case and has been shown to be significantly more accurate in virtually all domains (Ding et al. 2008; Rakthanmanon et al. 2013).”
Searching and Mining Trillions of Time Series Subsequences under Dynamic Time Warping (2012, background): “In addition, virtually all practitioners using DTW also constrain the warping path in a global sense by limiting how far it may stray from the diagonal [6][28].”