Finding the most unusual time series subsequence: algorithms and applications

Authors:
Eamonn Keogh;Jessica Lin;Sang-Hee Lee;Helga Van Herle
Affiliations:
University of California, Department of Computer Science and Engineering, Riverside, CA, USA;George Mason University, Department of Information and Software Engineering, Fairfax, VA, USA;University of California, Anthropology Department, Riverside, CA, USA;University of California, David Geffen School of Medicine, Los Angeles, CA, USA
Venue:
Knowledge and Information Systems
Year:
2006

Citing 0
Cited 13

Efficient query filtering for streaming time series with applications to semisupervised learning of time series classifiers

Knowledge and Information Systems
Mining temporal interval relational rules from temporal data

Journal of Systems and Software
Anomaly detection: A survey

ACM Computing Surveys (CSUR)
Network anomaly detection based on Eigen equation compression

Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining
Finding Structural Similarity in Time Series Data Using Bag-of-Patterns Representation

SSDBM 2009 Proceedings of the 21st International Conference on Scientific and Statistical Database Management
Finding the k-Most Abnormal Subgraphs from a Single Graph

DS '09 Proceedings of the 12th International Conference on Discovery Science
Evolutionary algorithms for the selection of time lags for time series forecasting by fuzzy inference systems

Neurocomputing
Unraveling complex temporal associations in cellular systems across multiple time-series microarray datasets

Journal of Biomedical Informatics
A review on time series data mining

Engineering Applications of Artificial Intelligence
Anomaly detection in monitoring sensor data for preventive maintenance

Expert Systems with Applications: An International Journal
Time-series data mining

ACM Computing Surveys (CSUR)
Rotation-invariant similarity in time series using bag-of-patterns representation

Journal of Intelligent Information Systems
Review: A review of novelty detection

Signal Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this work we introduce the new problem of finding time seriesdiscords. Time series discords are subsequences of longer time series that are maximally different to all the rest of the time series subsequences. They thus capture the sense of the most unusual subsequence within a time series. While discords have many uses for data mining, they are particularly attractive as anomaly detectors because they only require one intuitive parameter (the length of the subsequence) unlike most anomaly detection algorithms that typically require many parameters. While the brute force algorithm to discover time series discords is quadratic in the length of the time series, we show a simple algorithm that is three to four orders of magnitude faster than brute force, while guaranteed to produce identical results. We evaluate our work with a comprehensive set of experiments on diverse data sources including electrocardiograms, space telemetry, respiration physiology, anthropological and video datasets.