Exact indexing of dynamic time warping

By Eamonn J. Keogh, Chotirat (Ann) Ratanamahatana. Knowledge and Information Systems, 2002

2,764 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/7fb7364cf5f9bd48a9dd3a92a5243aae128afc24

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

Robust and fast similarity search for moving object trajectories (2005, methodology): “Unlike the pruning methods proposed for LCSS [36, 37] or DTW [19], these pruning methods do not require setting constraints on warping length (or matching region) between two trajectories, and therefore, offer users more flexibility.”
Analysis of Similarity Measures in Times Series Clustering for the Discovery of Building Energy Patterns (2013, background): “An interesting measure specially addressed to time series comparison is the Dynamic Time Warping (DTW) distance [21].”
Querying and mining of time series data: experimental comparison of representations and distance measures (2008, background): “Section 5 concludes the paper and discusses possible future extensions of the work.”
Making Time-Series Classification More Accurate Using Learned Constraints (2004, methodology): “We also report the accuracy at 10% warping window size since it is the number most researchers typically have been using in DTW researches [23][38][41].”
Three Myths about Dynamic Time Warping Data Mining (2005, methodology): “The steady flow of research papers on data mining with DTW became a torrent after it was shown that a simple lower bound allowed DTW to be indexed with no false dismissals [6].”
Toward accurate dynamic time warping in linear time and space (2007, neutral): “Indexing [9,14] uses lower-bounding functions to prune the number of times DTW is run for similarity search [17].”
Computing and Visualizing Dynamic Time Warping Alignments in R: The dtw Package (2009, methodology): “Larger problems may be addressed by approximate strategies, e.g., computing a preliminary alignment between downsampled time series (Salvador and Chan 2004); indexing (Keogh and Ratanamahatana 2005); or breaking one of the sequences into chunks and then iterating subsequence matches.”
Elasticity in Cloud Computing: What It Is, and What It Is Not (2013, methodology): “As an alternative to these metrics, the dynamic time warping (DTW) distance [7] can be used as a robust distance metric to capture the similarity between the demand and supply curves as well as to approximate the technical reaction time of the adaptation mechanism.”
Indexing spatio-temporal trajectories with Chebyshev polynomials (2004, background): “Some of the most famous distance notations are: • Dynamic Time Warping (DTW) [4, 19]: The idea is to use dynamic programming to construct the warping path in the distance matrix that minimizes the 5”
Logical-shapelets: an expressive primitive for time series classification (2011, methodology): “We also considered DTW with the Sakoe-Chiba band [15], as it has been shown to outperform classic DTW on many datasets [7].”
A Bag-of-Features Framework to Classify Time Series (2013, methodology): “Despite the fact that finding DTW distance without any modification on the algorithm is known to be computationally demanding for many applications [54], fast lower bounding function is used by [23] to prune the time series that cannot be the best match.”
The Complexities of Agent-Based Modeling Output Analysis (2015, methodology): “In addition to Richiardi's (2012) suggestions for robust TS comparison, we advocate the use of dynamic time warping (DTW) to address the above complication (Keogh & Ratanamahatana 2005).”
Experimental comparison of representation methods and distance measures for time series data (2010, methodology): “…is a very meaningful measure (Keogh et al. 2001b), and this also appears to be the current consensus in the literature (Cai and Ng 2004; Chen and Ng 2004; Chen et al. 2007a; Keogh 2002, 2006; Keogh et al. 2001a; Keogh and Ratanamahatana 2005; Ratanamahatana and Keogh 2005; Vlachos et al. 2006).”
A Survey on Graph Neural Networks for Time Series: Forecasting, Classification, Imputation, and Anomaly Detection (2023, methodology): “LB-SimTSC [97] further extends on SimTSC to improve the DTW preprocessing efﬁciency by employing the widely-used lower bound for DTW [163].”
On the Need for Time Series Data Mining Benchmarks: A Survey and Empirical Demonstration (2002, background): “It has been forcibly demonstrated that in some domains the Euclidean distance is a brittle distance measure, because it is very sensitive to small distortions in the time axis (Keogh, 2002; Berndt and Clifford, 1996; Yi et al., 1998).”