Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese Networks

In time-series data analysis, identifying anomalies is crucial for maintaining data integrity and ensuring accurate analyses and decision-making. Anomalies can compromise data quality and operational efficiency. The complexity of time-series data, with its temporal dependencies and potential non-sta...

Full description

Saved in:
Bibliographic Details
Main Authors: Ayan Chatterjee, Vajira Thambawita, Michael A. Riegler, Pal Halvorsen
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10966923/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850203790765981696
author Ayan Chatterjee
Vajira Thambawita
Michael A. Riegler
Pal Halvorsen
author_facet Ayan Chatterjee
Vajira Thambawita
Michael A. Riegler
Pal Halvorsen
author_sort Ayan Chatterjee
collection DOAJ
description In time-series data analysis, identifying anomalies is crucial for maintaining data integrity and ensuring accurate analyses and decision-making. Anomalies can compromise data quality and operational efficiency. The complexity of time-series data, with its temporal dependencies and potential non-stationarity, makes anomaly detection challenging but essential. Our research introduces ADSiamNet, a 1D Convolutional Neural Network-based Siamese network model for anomaly detection and rectification. ADSiamNet effectively identifies localized patterns in time-series data and smooths detected anomalies using a quantile-based technique. In tests with physical activity data from Actigraph watches and MOX2-5 sensors, ADSiamNet achieved accuracies of 98.65% and 85.0%, respectively, outperforming other supervised anomaly detection methods. The model uses a contrastive loss function to compare input sequences and adjusts network weights iteratively during training to recognize intricate patterns. Additionally, we evaluated various univariate time-series forecasting algorithms on datasets with and without anomalies. Results show that anomaly-smoothed data reduces forecasting errors, highlighting our approach’s effectiveness in enhancing time-series data analysis’s integrity and reliability. Future research will focus on multivariate time-series datasets.
format Article
id doaj-art-2b9eb3f3054e4688a9866ddefef3088e
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-2b9eb3f3054e4688a9866ddefef3088e2025-08-20T02:11:25ZengIEEEIEEE Access2169-35362025-01-0113709807100610.1109/ACCESS.2025.356137510966923Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese NetworksAyan Chatterjee0https://orcid.org/0000-0003-0407-7702Vajira Thambawita1https://orcid.org/0000-0001-6026-0929Michael A. Riegler2https://orcid.org/0000-0002-3153-2064Pal Halvorsen3https://orcid.org/0000-0003-2073-7029Department of Digital Technology, STIFTELSEN NILU, Kjeller, NorwayDepartment of Holistic Systems, Simula Metropolitan Center for Digital Engineering (SimulaMet), Oslo, NorwayDepartment of Information Technology, Oslo Metropolitan University, Oslo, NorwayDepartment of Information Technology, Oslo Metropolitan University, Oslo, NorwayIn time-series data analysis, identifying anomalies is crucial for maintaining data integrity and ensuring accurate analyses and decision-making. Anomalies can compromise data quality and operational efficiency. The complexity of time-series data, with its temporal dependencies and potential non-stationarity, makes anomaly detection challenging but essential. Our research introduces ADSiamNet, a 1D Convolutional Neural Network-based Siamese network model for anomaly detection and rectification. ADSiamNet effectively identifies localized patterns in time-series data and smooths detected anomalies using a quantile-based technique. In tests with physical activity data from Actigraph watches and MOX2-5 sensors, ADSiamNet achieved accuracies of 98.65% and 85.0%, respectively, outperforming other supervised anomaly detection methods. The model uses a contrastive loss function to compare input sequences and adjusts network weights iteratively during training to recognize intricate patterns. Additionally, we evaluated various univariate time-series forecasting algorithms on datasets with and without anomalies. Results show that anomaly-smoothed data reduces forecasting errors, highlighting our approach’s effectiveness in enhancing time-series data analysis’s integrity and reliability. Future research will focus on multivariate time-series datasets.https://ieeexplore.ieee.org/document/10966923/Univariate time-seriesanomaly detectionsiamese neural networks1D CNNsquantile smoothingpredictive modeling
spellingShingle Ayan Chatterjee
Vajira Thambawita
Michael A. Riegler
Pal Halvorsen
Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese Networks
IEEE Access
Univariate time-series
anomaly detection
siamese neural networks
1D CNNs
quantile smoothing
predictive modeling
title Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese Networks
title_full Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese Networks
title_fullStr Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese Networks
title_full_unstemmed Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese Networks
title_short Supervised Anomaly Detection in Univariate Time-Series Using 1D Convolutional Siamese Networks
title_sort supervised anomaly detection in univariate time series using 1d convolutional siamese networks
topic Univariate time-series
anomaly detection
siamese neural networks
1D CNNs
quantile smoothing
predictive modeling
url https://ieeexplore.ieee.org/document/10966923/
work_keys_str_mv AT ayanchatterjee supervisedanomalydetectioninunivariatetimeseriesusing1dconvolutionalsiamesenetworks
AT vajirathambawita supervisedanomalydetectioninunivariatetimeseriesusing1dconvolutionalsiamesenetworks
AT michaelariegler supervisedanomalydetectioninunivariatetimeseriesusing1dconvolutionalsiamesenetworks
AT palhalvorsen supervisedanomalydetectioninunivariatetimeseriesusing1dconvolutionalsiamesenetworks