Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data

Sreelekha Guggilam; Varun Chandola; Abani K. Patra

doi:10.1007/978-3-031-08751-6_10

Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data

Sreelekha Guggilam, Varun Chandola, Abani K. Patra

GeoAI

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Anomaly detection for time series data is often aimed at identifying extreme behaviors within an individual time series. However, identifying extreme trends relative to a collection of other time series is of significant interest, like in the fields of public health policy, social justice and pandemic propagation. We propose an algorithm that can scale to large collections of time series data using the concepts from the theory of large deviations. Exploiting the ability of the algorithm to scale to high-dimensional data, we propose an online anomaly detection method to identify anomalies in a collection of multivariate time series. We demonstrate the applicability of the proposed Large Deviations Anomaly Detection (LAD) algorithm in identifying counties in the United States with anomalous trends in terms of COVID-19 related cases and deaths. Several of the identified anomalous counties correlate with counties with documented poor response to the COVID pandemic.

Original language	English
Title of host publication	Computational Science - ICCS 2022, 22nd International Conference, Proceedings
Editors	Derek Groen, Clélia de Mulatier, Valeria V. Krzhizhanovskaya, Peter M.A. Sloot, Maciej Paszynski, Jack J. Dongarra
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	133-149
Number of pages	17
ISBN (Print)	9783031087509
DOIs	https://doi.org/10.1007/978-3-031-08751-6_10
State	Published - 2022
Event	22nd Annual International Conference on Computational Science, ICCS 2022 - London, United Kingdom Duration: Jun 21 2022 → Jun 23 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13350 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	22nd Annual International Conference on Computational Science, ICCS 2022
Country/Territory	United Kingdom
City	London
Period	06/21/22 → 06/23/22

Funding

Acknowledgements. The authors would like to acknowledge University at Buffalo Center for Computational Research for computing resources and financial support of the National Science Foundation Grant numbers NSF/OAC 1339765 and NSF/DMS 1621853.

Keywords

Anomaly detection
High-dimensional data
Large deviations
Multivariate time series
Time series database

Access to Document

10.1007/978-3-031-08751-6_10

Cite this

Guggilam, S., Chandola, V., & Patra, A. K. (2022). Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data. In D. Groen, C. de Mulatier, V. V. Krzhizhanovskaya, P. M. A. Sloot, M. Paszynski, & J. J. Dongarra (Eds.), Computational Science - ICCS 2022, 22nd International Conference, Proceedings (pp. 133-149). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13350 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-08751-6_10

Guggilam, Sreelekha ; Chandola, Varun ; Patra, Abani K. / Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle : An Application to COVID-19 Data. Computational Science - ICCS 2022, 22nd International Conference, Proceedings. editor / Derek Groen ; Clélia de Mulatier ; Valeria V. Krzhizhanovskaya ; Peter M.A. Sloot ; Maciej Paszynski ; Jack J. Dongarra. Springer Science and Business Media Deutschland GmbH, 2022. pp. 133-149 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{a70ebd785b614aa19a955f50025733b1,

title = "Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data",

abstract = "Anomaly detection for time series data is often aimed at identifying extreme behaviors within an individual time series. However, identifying extreme trends relative to a collection of other time series is of significant interest, like in the fields of public health policy, social justice and pandemic propagation. We propose an algorithm that can scale to large collections of time series data using the concepts from the theory of large deviations. Exploiting the ability of the algorithm to scale to high-dimensional data, we propose an online anomaly detection method to identify anomalies in a collection of multivariate time series. We demonstrate the applicability of the proposed Large Deviations Anomaly Detection (LAD) algorithm in identifying counties in the United States with anomalous trends in terms of COVID-19 related cases and deaths. Several of the identified anomalous counties correlate with counties with documented poor response to the COVID pandemic.",

keywords = "Anomaly detection, High-dimensional data, Large deviations, Multivariate time series, Time series database",

author = "Sreelekha Guggilam and Varun Chandola and Patra, {Abani K.}",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 22nd Annual International Conference on Computational Science, ICCS 2022 ; Conference date: 21-06-2022 Through 23-06-2022",

year = "2022",

doi = "10.1007/978-3-031-08751-6_10",

language = "English",

isbn = "9783031087509",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "133--149",

editor = "Derek Groen and {de Mulatier}, Cl{\'e}lia and Krzhizhanovskaya, {Valeria V.} and Sloot, {Peter M.A.} and Maciej Paszynski and Dongarra, {Jack J.}",

booktitle = "Computational Science - ICCS 2022, 22nd International Conference, Proceedings",

}

Guggilam, S, Chandola, V & Patra, AK 2022, Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data. in D Groen, C de Mulatier, VV Krzhizhanovskaya, PMA Sloot, M Paszynski & JJ Dongarra (eds), Computational Science - ICCS 2022, 22nd International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13350 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 133-149, 22nd Annual International Conference on Computational Science, ICCS 2022, London, United Kingdom, 06/21/22. https://doi.org/10.1007/978-3-031-08751-6_10

Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data. / Guggilam, Sreelekha; Chandola, Varun; Patra, Abani K.
Computational Science - ICCS 2022, 22nd International Conference, Proceedings. ed. / Derek Groen; Clélia de Mulatier; Valeria V. Krzhizhanovskaya; Peter M.A. Sloot; Maciej Paszynski; Jack J. Dongarra. Springer Science and Business Media Deutschland GmbH, 2022. p. 133-149 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13350 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle

T2 - 22nd Annual International Conference on Computational Science, ICCS 2022

AU - Guggilam, Sreelekha

AU - Chandola, Varun

AU - Patra, Abani K.

PY - 2022

Y1 - 2022

N2 - Anomaly detection for time series data is often aimed at identifying extreme behaviors within an individual time series. However, identifying extreme trends relative to a collection of other time series is of significant interest, like in the fields of public health policy, social justice and pandemic propagation. We propose an algorithm that can scale to large collections of time series data using the concepts from the theory of large deviations. Exploiting the ability of the algorithm to scale to high-dimensional data, we propose an online anomaly detection method to identify anomalies in a collection of multivariate time series. We demonstrate the applicability of the proposed Large Deviations Anomaly Detection (LAD) algorithm in identifying counties in the United States with anomalous trends in terms of COVID-19 related cases and deaths. Several of the identified anomalous counties correlate with counties with documented poor response to the COVID pandemic.

AB - Anomaly detection for time series data is often aimed at identifying extreme behaviors within an individual time series. However, identifying extreme trends relative to a collection of other time series is of significant interest, like in the fields of public health policy, social justice and pandemic propagation. We propose an algorithm that can scale to large collections of time series data using the concepts from the theory of large deviations. Exploiting the ability of the algorithm to scale to high-dimensional data, we propose an online anomaly detection method to identify anomalies in a collection of multivariate time series. We demonstrate the applicability of the proposed Large Deviations Anomaly Detection (LAD) algorithm in identifying counties in the United States with anomalous trends in terms of COVID-19 related cases and deaths. Several of the identified anomalous counties correlate with counties with documented poor response to the COVID pandemic.

KW - Anomaly detection

KW - High-dimensional data

KW - Large deviations

KW - Multivariate time series

KW - Time series database

UR - http://www.scopus.com/inward/record.url?scp=85134352191&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-08751-6_10

DO - 10.1007/978-3-031-08751-6_10

M3 - Conference contribution

AN - SCOPUS:85134352191

SN - 9783031087509

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 133

EP - 149

BT - Computational Science - ICCS 2022, 22nd International Conference, Proceedings

A2 - Groen, Derek

A2 - de Mulatier, Clélia

A2 - Krzhizhanovskaya, Valeria V.

A2 - Sloot, Peter M.A.

A2 - Paszynski, Maciej

A2 - Dongarra, Jack J.

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 21 June 2022 through 23 June 2022

ER -

Guggilam S, Chandola V, Patra AK. Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data. In Groen D, de Mulatier C, Krzhizhanovskaya VV, Sloot PMA, Paszynski M, Dongarra JJ, editors, Computational Science - ICCS 2022, 22nd International Conference, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 133-149. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-08751-6_10