ECG Anomaly Detection using Time Series Deep Learning Algorithms

Mohammed Ayden  Omar; Layth  Hammood

Mohammed Ayden Omar Ministry of Planning, Kirkuk Planning Director, Kirkuk, Iraq
Layth Hammood Presidency of University of Kirkuk, University of Kirkuk, Kirkuk, Iraq

Keywords: Timeseries, Variational Inference, VAE, Anomaly Detection

Abstract

This paper implements a version of a VAE for anomaly detection known as VELC. VELC uses a re-encoder at the end of the decoder. The results are evaluated with several versions of vanilla VAEs containing only an encoder and decoder. Initially, we built models that yield reliable results for the ECG dataset, and thereafter we used those models with the regular passenger cars dataset provided by the instructors of the course. We measured the performance of the models using the AUC. Thresholding different values of the mean squared error score between the actual and reconstructed time series calculated the AUC curve. Next, the best results were obtained with a bidirectional LSTM with a re-encoder for a value of 0.997. The regular passenger cars data yielded poor results for this method of classification, which motivated a different approach thresholding the Lilliefors test statistic of the residuals. This provided better results, with the best AUC value of 0.754 for a bidirectional GRU without re-encoding.

References

G. Zhu, H. Zhao, H. Liu, and H. Sun, "A novel LSTM-GAN algorithm for time series anomaly detection," in Proc. 2019 Prognostics and System Health Management Conference (PHM-Qingdao), Qingdao, China, Oct. 2019, pp. 1-6.

S. Schmidl, P. Wenig, and T. Papenbrock, "Anomaly detection in time series: a comprehensive evaluation," Proc. VLDB Endowment, vol. 15, no. 9, pp. 1779-1797, 2022.

P. Liu, X. Sun, Y. Han, Z. He, W. Zhang, and C. Wu, "Arrhythmia classification of LSTM autoencoder based on time series anomaly detection," Biomed. Signal Process. Control, vol. 71, p. 103228, 2022.

M. Roy, S. Majumder, A. Halder, and U. Biswas, "ECG-NET: A deep LSTM autoencoder for detecting anomalous ECG," Eng. Appl. Artif. Intell., vol. 124, p. 106484, 2023.

L. Goldberger et al., "PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals," Circulation, vol. 101, no. 23, pp. e215-e220, 2000.

M. Thill, W. Konen, H. Wang, and T. Bäck, "Temporal convolutional autoencoder for unsupervised anomaly detection in time series," Appl. Soft Comput., vol. 112, p. 107751, 2021.

A. Pol, V. Berger, C. Germain, G. Cerminara, and M. Pierini, "Anomaly detection with conditional variational autoencoders," in Proc. 2019 18th IEEE Int. Conf. Machine Learning Applications (ICMLA), pp. 1651-1657, 2019.

J. An and S. Cho, "Variational autoencoder based anomaly detection using reconstruction probability," Special Lecture on IE, vol. 2, no. 1, pp. 1-18, 2015.

D. Gong et al., "Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection," in Proc. IEEE/CVF Int. Conf. Computer Vision, 2019, pp. 1705-1714.

C. Zhang, S. Li, H. Zhang, and Y. Chen, "VELC: A new variational autoencoder based model for time series anomaly detection," arXiv preprint arXiv:1907.01702, 2019.

K. Dutta, R. Lenka, S. R. Nayak, A. Khandual, and A. K. Bhoi, "MED-NET: A novel approach to ECG anomaly detection using LSTM auto-encoders," Int. J. Comput. Appl. Technol., vol. 65, no. 4, pp. 343-357, 2021.

G. Li and J. J. Jung, "Deep learning for anomaly detection in multivariate time series: Approaches, applications, and challenges," Inf. Fusion, vol. 91, pp. 93-102, 2023.

S. Maya, K. Ueno, and T. Nishikawa, "dLSTM: A new approach for anomaly detection using deep learning with delayed prediction," Int. J. Data Sci. Anal., vol. 8, pp. 137-164, 2019.