2026-09-20T09:24:20Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1446712022-08-30T03:01:54Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Veeramachaneni, Kalyan
Wong, Lawrence C.
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2022-08-29T16:03:41Z
2022-08-29T16:03:41Z
2022-05
2022-05-27T16:19:10.928Z
https://hdl.handle.net/1721.1/144671
Anomaly detection on time series data is increasingly common across various industrial domains that require monitoring metrics to prevent potential accidents and economic losses. The complications of anomaly detection revolve around a scarcity of labeled data and the need to learn temporal correlations between multiple variables. Most successful unsupervised methods either use single-timestamp prediction or reconstruct entire time series. However, these methods are not mutually exclusive and can each offer complementary perspectives. This work first explores the successes and limitations of prediction-based and reconstruction-based methods. Next, it compares the effect of attention-based architectures with LSTM-based architectures on existing models. Finally, this research proposes a novel autoencoder architecture capable of producing bi-directional predictions while simultaneously reconstructing the original time series by optimizing a joint objective function. An ablation study using a mixture of prediction and reconstruction errors demonstrates that this simple architecture outperforms other state-of-the-art models for anomaly detection on both univariate and multivariate time series.
M.Eng.
Massachusetts Institute of Technology
In Copyright - Educational Use Permitted
Copyright MIT
http://rightsstatements.org/page/InC-EDU/1.0/
Time Series Anomaly Detection using Prediction-Reconstruction Mixture Errors
Thesis
application/pdf
Master
Master of Engineering in Computer Science and Molecular Biology
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="f534a51d-89d0-41ca-9ac0-9d077e015a6d">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Time Series Anomaly Detection using Prediction-Reconstruction Mixture Errors</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2022-05</PublicationDate>
<Authors>
<Author>
<DisplayName>Wong, Lawrence C.</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>http://rightsstatements.org/page/InC-EDU/1.0/</License>
<Abstract>Anomaly detection on time series data is increasingly common across various industrial domains that require monitoring metrics to prevent potential accidents and economic losses. The complications of anomaly detection revolve around a scarcity of labeled data and the need to learn temporal correlations between multiple variables. Most successful unsupervised methods either use single-timestamp prediction or reconstruct entire time series. However, these methods are not mutually exclusive and can each offer complementary perspectives. This work first explores the successes and limitations of prediction-based and reconstruction-based methods. Next, it compares the effect of attention-based architectures with LSTM-based architectures on existing models. Finally, this research proposes a novel autoencoder architecture capable of producing bi-directional predictions while simultaneously reconstructing the original time series by optimizing a joint objective function. An ablation study using a mixture of prediction and reconstruction errors demonstrates that this simple architecture outperforms other state-of-the-art models for anomaly detection on both univariate and multivariate time series.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>