Temporal convolutional autoencoder for unsupervised anomaly detection in time series

• Published in: Applied soft computing Vol. 112; p. 107751
• Main Authors: Thill, Markus, Konen, Wolfgang, Wang, Hao, Bäck, Thomas
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.11.2021
• Subjects: Anomaly detection; Autoencoder; Deep learning; Mahalanobis distance; TCN; Deep learning; Autoencoder; Anomaly detection; TCN; Mahalanobis distance
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2021.107751

Learning temporal patterns in time series remains a challenging task up until today. Particularly for anomaly detection in time series, it is essential to learn the underlying structure of a system’s normal behavior. Periodic or quasiperiodic signals with complex temporal patterns make the problem even more challenging: Anomalies may be a hard-to-detect deviation from the normal recurring pattern. In this paper, we present TCN-AE, a temporal convolutional network autoencoder based on dilated convolutions. Contrary to many other anomaly detection algorithms, TCN-AE is trained in an unsupervised manner. The algorithm demonstrates its efficacy on a comprehensive real-world anomaly benchmark comprising electrocardiogram (ECG) recordings of patients with cardiac arrhythmia. TCN-AE significantly outperforms several other unsupervised state-of-the-art anomaly detection algorithms. Moreover, we investigate the contribution of the individual enhancements and show that each new ingredient improves the overall performance on the investigated benchmark. •Novel Temporal Convolutional Network Auto-Encoder for time series anomaly detection.•Unsupervised learning of time series representations.•High performance on real-world anomaly detection task containing electrocardiograms.•The presented algorithm is also computationally very efficient.