Time-Series Classification with COTE: The Collective of Transformation-Based Ensembles

By Anthony J. Bagnall, Jason Lines, Jon Hills, Aaron Bostrom. IEEE Transactions on Knowledge and Data Engineering, 2015

470 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Signal Processing

View Paper: https://www.semanticscholar.org/paper/be9892e207007d5f4c3e23133ea03daef64b87b0

How researchers see this paper

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

In Economics and Econometrics

1. Geoffrey I. Webb
2. Anthony J. Bagnall
3. Franccois Petitjean
4. Jason Lines
5. Aaron Bostrom
6. Charlotte Pelletier
7. Pierre-Alain Muller
8. Jonathan Weber

In Signal Processing

1. Geoffrey I. Webb
2. Franccois Petitjean
3. Anthony J. Bagnall
4. Charlotte Pelletier
5. Jason Lines
6. Daniel F. Schmidt
7. Aaron Bostrom
8. Ben D. Fulcher

Citations in the corpus

• Temporal Convolutional Neural Network for the Classification of Satellite Image Time Series (2018, background): “Collective Of Transformation-based Ensembles: Both Collective Of Transformation-based Ensembles (COTE) algorithms – Flat-COTE [77] and the more recent Hierarchical Vote of COTE (HIVE-COTE) [78] – are meta-ensemble algorithms that include a set of classifiers working with different representations of”
• catch22: CAnonical Time-series CHaracteristics (2019, methodology): “Note that hybrid methods such as COTE (Bagnall et al. 2016), which are not limited to a single time-series representation but can adapt to the problem at hand, consistently outperform both the shape-based existing classifiers and our features at the price of a much higher computational effort.”
• Deep learning for time series classification: a review (2018, methodology): “Despite the fact that a huge number of parameters risks overfitting (Zhang et al., 2017) the relatively small train set in the UCR/UEA archive, our experiments showed that not only DNNs are able to significantly outperform the NN-DTW, but are also able to achieve results that are not significantly d”
• Time series classification from scratch with deep neural networks: A strong baseline (2016, background): “Shapelet ensemble (SE) [7] produces the classifiers through the shapelet transform in conjunction with a heterogeneous ensemble.”
• InceptionTime: Finding AlexNet for time series classification (2019, background): “More recently, Bagnall et al. (2016) noted that there is no single time series transformation technique (such as shapelets or SFA) that significantly dominates the others, showing that constructing an ensemble of different classifiers over different time series representations, called COTE, will…”
• Unsupervised Scalable Representation Learning for Multivariate Time Series (2019, background): “(2017)4: COTE (Bagnall et al., 2015), ST (Bostrom & Bagnall, 2015), BOSS (Schäfer, 2015) and EE (Lines & Bagnall, 2015), on the first 85 datasets of the archive5.”
• The great time series classification bake off: a review and experimental evaluation of recent algorithmic advances (2016, methodology): “Following Bagnall et al. (2015) and Bostrom and Bagnall (2015)we construct a classifier from this dataset using aweighted ensemble of standard classifiers.”