A Comparative Study of Deep Learning Architectures for Beat-Level Arrhythmia Classification using the MIT-BIH Database

Ceena Mathews

doi:10.14445/22312803/IJCTT-V74I5P102

Research Article | Open Access | Download PDF

Volume 74 | Issue 5 | Year 2026 | Article Id. IJCTT-V74I5P102 | DOI : https://doi.org/10.14445/22312803/IJCTT-V74I5P102

A Comparative Study of Deep Learning Architectures for Beat-Level Arrhythmia Classification using the MIT-BIH Database

Ceena Mathews

Received	Revised	Accepted	Published
16 Mar 2026	21 Apr 2026	12 May 2026	28 May 2026

Citation :

Ceena Mathews, "A Comparative Study of Deep Learning Architectures for Beat-Level Arrhythmia Classification using the MIT-BIH Database," International Journal of Computer Trends and Technology (IJCTT), vol. 74, no. 5, pp. 10-14, 2026. Crossref, https://doi.org/10.14445/22312803/IJCTT-V74I5P102

Abstract

Cardiovascular diseases continue to be a major cause of mortality worldwide, and among these, cardiac arrhythmias still remain challenging to diagnose accurately. Electrocardiogram (ECG) analysis is a non-invasive tool for the detection of arrhythmias. Manual interpretation of ECG signals can be time-consuming and often varies between clinicians. In this work, six deep learning architectures, ResNet1D-34, InceptionTime, transformer, attention-BiLSTM, attention-aware pooling, and convolution-enhanced transformer, are implemented for beat-level arrhythmia classification using the MIT-BIH arrhythmia database under AAMI class grouping (Normal, Superventricular, Ventricular, Other). All these models are evaluated using macro F1-score to better account for class imbalance, and class-wise evaluation is also assessed to better predict the model performance. The results show that convolution-based models tend to perform more consistently and achieve better balance across classes, while attention-based models struggle with classes that have similar waveform patterns, particularly supraventricular beats.

Keywords

Attention, Convolution, ECG beats, InceptionTime, LSTM, Transformer.

References

[1] Antônio H. Ribeiro et al., “Automatic Diagnosis of the 12-Lead ECG Using a Deep Neural Network,” Nature Communications, vol. 11, no. 1, pp. 1-9, 2020.
[CrossRef] [Google Scholar] [Publisher Link]

[2] Shenda Hong et al., “Opportunities and Challenges of Deep Learning Methods for Electrocardiogram Data: A Systematic Review,” Computers in Biology and Medicine, vol. 122, 2020.
[CrossRef] [Google Scholar] [Publisher Link]

[3] Yaqoob Ansari et al., “Deep Learning for ECG Arrhythmia Detection and Classification: An Overview of Progress for Period 2017–2023,” Frontiers in Physiology, vol. 14, 2023.
[CrossRef] [Google Scholar] [Publisher Link]

[4] Hassan Ismail Fawaz et al., “InceptionTime: Finding AlexNet for Time Series Classification,” Data Mining and Knowledge Discovery, vol. 34, no. 6, pp. 1936-1962, 2020.
[CrossRef] [Google Scholar] [Publisher Link]

[5] Shahab Ul Hassan et al., “Classification of Cardiac Arrhythmia Using a Convolutional Neural Network and Bi-Directional Long Short-Term Memory,” Digital Health, vol. 8, 2022.
[CrossRef] [Google Scholar] [Publisher Link]

[6] Mingfeng Jiang et al., “HADLN: Hybrid Attention-Based Deep Learning Network for Automated Arrhythmia Classification,” Frontiers in Physiology, vol. 12, 2021.
[CrossRef] [Google Scholar] [Publisher Link]

[7] Taymaz Akan, Sait Alp, and Mohammad Alfrad Nobel Bhuiyan, “ECGformer: Leveraging Transformer for ECG Heartbeat Arrhythmia Classification,” 2023 International Conference on Computational Science and Computational Intelligence (CSCI), Las Vegas, NV, USA, pp. 1412-1417, 2023.
[CrossRef] [Google Scholar] [Publisher Link]

[8] Hany El-Ghaish, and Emadeldeen Eldele, “ECGTransForm: Empowering Adaptive ECG Arrhythmia Classification Framework with Bidirectional Transformer,” Biomedical Signal Processing and Control, vol. 89, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[9] Rui Hu, Jie Chen, and Li Zhou, “A Transformer-Based Deep Neural Network for Arrhythmia Detection Using Continuous ECG Signals,” Computers in Biology and Medicine, vol. 144, 2022.
[CrossRef] [Google Scholar] [Publisher Link]

[10] Donghyeon Kim et al., “A Novel Hybrid CNN-Transformer Model for Arrhythmia Detection without R-Peak Identification using Stockwell Transform,” Scientific Reports, vol. 15, no. 1, 2025.
[CrossRef] [Google Scholar] [Publisher Link]

[11] G.B. Moody, and R.G. Mark, “The Impact of the MIT-BIH Arrhythmia Database,” IEEE Engineering in Medicine and Biology Magazine, vol. 20, no. 3, pp. 45-50, 2001.
[CrossRef] [Google Scholar] [Publisher Link]