Explainable AI-Powered ECG Anomaly Detection Using SHAP and LSTM Autoencoders
Engineering, Technology & Applied Science Research · 06.06.2026 · 16 (3) · 34998-35005
DOI: 10.48084/etasr.17734 ↗Kurz gefasst
Fachartikel in der Zeitschrift Engineering, Technology & Applied Science Research, online erschienen am 06.06.2026. Verfasst von K. Nayana und 1 weiteren Autor:innen. Die Arbeit nennt 14 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Cardiovascular diseases are the leading cause of mortality, underscoring the need for accurate and early detection of cardiac abnormalities. This study proposes a Long Short-Term Memory (LSTM) autoencoder framework for automated Electrocardiogram (ECG) anomaly detection, integrated with SHapley Additive exPlanations (SHAP) to ensure transparent and clinically interpretable decision-making. Using the ECG5000 dataset, the model learns the intrinsic characteristics of normal cardiac cycles and identifies deviations through reconstruction error analysis. The proposed framework achieved an accuracy of 97.6%, a precision of 95.8%, and an F1-score 96.7%, outperforming traditional machine-learning baselines and earlier deep-learning approaches. SHAP-based visualization highlights the specific temporal segments influencing anomaly predictions, enhancing clinical trust and applicability. This work demonstrates a robust, explainable, and efficient approach, suitable for real-time cardiac monitoring.
Zitieren
APA
Nayana, K., & Vinay, S. (2026). Explainable AI-Powered ECG Anomaly Detection Using SHAP and LSTM Autoencoders. Engineering, Technology & Applied Science Research, 16(3), 34998-35005. https://doi.org/10.48084/etasr.17734
Vancouver
Nayana K, Vinay S. Explainable AI-Powered ECG Anomaly Detection Using SHAP and LSTM Autoencoders. Engineering, Technology & Applied Science Research. 2026;16(3):34998-35005. doi:10.48084/etasr.17734
BibTeX
@article{Nayana2026,
author = {Nayana, K. and Vinay, S.},
title = {Explainable AI-Powered ECG Anomaly Detection Using SHAP and LSTM Autoencoders},
journal = {Engineering, Technology & Applied Science Research},
year = {2026},
volume = {16},
number = {3},
pages = {34998--35005},
doi = {10.48084/etasr.17734}
}
EUVIMED