Standard

Autoencoder and Kernel Density Estimation Based Approach for Time Series Anomaly Detection. / Аржа, Антон Насерович; Корхов, Владимир Владиславович.

Computational Science and Its Applications – ICCSA 2025 Workshops. 2025. стр. 249–263 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Том 15894).

Результаты исследований: Публикации в книгах, отчётах, сборниках, трудах конференцийстатья в сборнике материалов конференцииРецензирование

Harvard

Аржа, АН & Корхов, ВВ 2025, Autoencoder and Kernel Density Estimation Based Approach for Time Series Anomaly Detection. в Computational Science and Its Applications – ICCSA 2025 Workshops. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Том. 15894, стр. 249–263, 25th International Conference on Computational Science and Its Applications, ICCSA 2025, Стамбул, Турция, 30/06/25. https://doi.org/10.1007/978-3-031-97648-3_17

APA

Аржа, А. Н., & Корхов, В. В. (2025). Autoencoder and Kernel Density Estimation Based Approach for Time Series Anomaly Detection. в Computational Science and Its Applications – ICCSA 2025 Workshops (стр. 249–263). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Том 15894). https://doi.org/10.1007/978-3-031-97648-3_17

Vancouver

Аржа АН, Корхов ВВ. Autoencoder and Kernel Density Estimation Based Approach for Time Series Anomaly Detection. в Computational Science and Its Applications – ICCSA 2025 Workshops. 2025. стр. 249–263. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-031-97648-3_17

Author

Аржа, Антон Насерович ; Корхов, Владимир Владиславович. / Autoencoder and Kernel Density Estimation Based Approach for Time Series Anomaly Detection. Computational Science and Its Applications – ICCSA 2025 Workshops. 2025. стр. 249–263 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

BibTeX

@inproceedings{f1db92de51004ad784ca05067e540be0,
title = "Autoencoder and Kernel Density Estimation Based Approach for Time Series Anomaly Detection",
abstract = "The detection of anomalies in time series is a critical task across various domains including time series generated by a diversity of distributed systems. Despite extensive research, no single algorithm is the undisputed leader in all respects. This paper introduces KDEAE, a novel unsupervised approach for univariate time series anomaly detection. This method combines autoencoders (AE) and kernel density estimation (KDE). The proposed approach segments time series into non-overlapping segments, encodes them using convolutional AE and constructs anomaly scores based on their density obtained from KDE. The method is evaluated on GutenTAG dataset collection using TimeEval framework and compared with 25 algorithms of the same category. The results show that KDEAE outperform 40% algorithms by ROC AUC and 60% by PR AUC scores and 60% by runtime. The paper concludes with suggestions for further research to enhance the method's performance and extend it to multivariate time series.",
keywords = "Anomaly Detection, Autoencoders, Kernel Density Estimation, Time Series",
author = "Аржа, {Антон Насерович} and Корхов, {Владимир Владиславович}",
year = "2025",
month = jun,
day = "28",
doi = "10.1007/978-3-031-97648-3_17",
language = "English",
isbn = "9783031976476",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Nature",
pages = "249–263",
booktitle = "Computational Science and Its Applications – ICCSA 2025 Workshops",
note = "Computational Science and Its Applications – ICCSA 2025 Workshops ; Conference date: 30-06-2025 Through 03-07-2025",
url = "http://iccsa.org",

}

RIS

TY - GEN

T1 - Autoencoder and Kernel Density Estimation Based Approach for Time Series Anomaly Detection

AU - Аржа, Антон Насерович

AU - Корхов, Владимир Владиславович

PY - 2025/6/28

Y1 - 2025/6/28

N2 - The detection of anomalies in time series is a critical task across various domains including time series generated by a diversity of distributed systems. Despite extensive research, no single algorithm is the undisputed leader in all respects. This paper introduces KDEAE, a novel unsupervised approach for univariate time series anomaly detection. This method combines autoencoders (AE) and kernel density estimation (KDE). The proposed approach segments time series into non-overlapping segments, encodes them using convolutional AE and constructs anomaly scores based on their density obtained from KDE. The method is evaluated on GutenTAG dataset collection using TimeEval framework and compared with 25 algorithms of the same category. The results show that KDEAE outperform 40% algorithms by ROC AUC and 60% by PR AUC scores and 60% by runtime. The paper concludes with suggestions for further research to enhance the method's performance and extend it to multivariate time series.

AB - The detection of anomalies in time series is a critical task across various domains including time series generated by a diversity of distributed systems. Despite extensive research, no single algorithm is the undisputed leader in all respects. This paper introduces KDEAE, a novel unsupervised approach for univariate time series anomaly detection. This method combines autoencoders (AE) and kernel density estimation (KDE). The proposed approach segments time series into non-overlapping segments, encodes them using convolutional AE and constructs anomaly scores based on their density obtained from KDE. The method is evaluated on GutenTAG dataset collection using TimeEval framework and compared with 25 algorithms of the same category. The results show that KDEAE outperform 40% algorithms by ROC AUC and 60% by PR AUC scores and 60% by runtime. The paper concludes with suggestions for further research to enhance the method's performance and extend it to multivariate time series.

KW - Anomaly Detection

KW - Autoencoders

KW - Kernel Density Estimation

KW - Time Series

UR - https://www.mendeley.com/catalogue/9292d4e6-8222-3c64-bd2d-48b15a173e85/

U2 - 10.1007/978-3-031-97648-3_17

DO - 10.1007/978-3-031-97648-3_17

M3 - Conference contribution

SN - 9783031976476

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

SP - 249

EP - 263

BT - Computational Science and Its Applications – ICCSA 2025 Workshops

T2 - Computational Science and Its Applications – ICCSA 2025 Workshops

Y2 - 30 June 2025 through 3 July 2025

ER -

ID: 138833380