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

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-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 languageEnglish
Title of host publicationComputational Science - ICCS 2022, 22nd International Conference, Proceedings
EditorsDerek Groen, Clélia de Mulatier, Valeria V. Krzhizhanovskaya, Peter M.A. Sloot, Maciej Paszynski, Jack J. Dongarra
PublisherSpringer Science and Business Media Deutschland GmbH
Pages133-149
Number of pages17
ISBN (Print)9783031087509
DOIs
StatePublished - 2022
Externally publishedYes
Event22nd Annual International Conference on Computational Science, ICCS 2022 - London, United Kingdom
Duration: Jun 21 2022Jun 23 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13350 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd Annual International Conference on Computational Science, ICCS 2022
Country/TerritoryUnited Kingdom
CityLondon
Period06/21/2206/23/22

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.

FundersFunder number
University at Buffalo Center for Computational Research
National Science Foundation
Division of Mathematical Sciences1621853
Ohio Arts Council1339765

    Keywords

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

    Fingerprint

    Dive into the research topics of 'Classifying Anomalous Members in a Collection of Multivariate Time Series Data Using Large Deviations Principle: An Application to COVID-19 Data'. Together they form a unique fingerprint.

    Cite this