ECG Signal Classification Using Multi-Resolution Wavelet Features and Gradient Boosting Frameworks

Jinge Bai

doi:10.62051/ijcsit.v6n1.14

Authors

Jinge Bai

DOI:

https://doi.org/10.62051/ijcsit.v6n1.14

Keywords:

Machine Learning, ECG Classification, Ensemble Methods, Feature Extraction

Abstract

Accurate and automated classification of Electrocardiogram (ECG) signals is critical for early diagnosis of cardiovascular diseases. In this study, we propose a hybrid machine learning framework that combines multi-resolution analysis through Discrete Wavelet Transform (DWT) with powerful ensemble learning models, specifically Gradient Boosting and LightGBM classifiers. Multi-scale time-frequency features were extracted using diverse wavelet families, and were further processed through feature standardization and dimensionality reduction techniques including Principal Component Analysis (PCA) and t-Distributed Stochastic Neighbor Embedding (t-SNE). Extensive experiments demonstrate that LightGBM achieves superior classification performance, reaching an accuracy of 98.41%, while Gradient Boosting attains an accuracy of 95.24%. The results highlight the effectiveness of combining handcrafted feature engineering from the wavelet domain with advanced boosting techniques for robust ECG signal categorization. This framework offers a promising tool for developing intelligent diagnostic support systems in clinical practice.

Downloads

Download data is not yet available.

References

[1] Wu, B. (2019). Research on ECG-based identity recognition and arrhythmia classification methods [Doctoral dissertation, Shandong University].

[2] Wang, T., Lu, C., Sun, Y. N., et al. (2021). Automatic ECG classification using continuous wavelet transform and convolutional neural network. Entropy, 23(1). https://doi.org/10.3390/e23010119

[3] Ma, J. W., & Liu, S. P. (2018). A review of ECG signal recognition and classification algorithms. Journal of Chongqing University of Technology (Natural Science), 32(12), 122–128.

[4] Hou, X. Q., Tong, Z. Y., & Liu, X. W. (2020). Research on QRS feature extraction algorithm based on improved wavelet transform. Modern Electronics Technique, 43(13), 57–61.

[5] Wang, C. C., Peng, Y., Liao, Y., et al. (2020). Research on ECG signal denoising algorithm based on improved wavelet threshold function. Electronic Technology & Software Engineering, 1, 74–75.

[6] Xu, B. X. (2020). Application of ECG signal denoising and T-wave automatic detection based on computer deep learning. Journal of Physics: Conference Series, 1578(1). https://doi.org/10.1088/1742-6596/1578/1/012007

[7] Song, L., Meng, Q. J., Zhang, G. Y., et al. (2010). Automatic ECG classification method based on waveform features and SVM. Chinese Journal of Medical Physics, 27(4), 2043–2046.

[8] Wu, Z. Y., Ding, X. Q., Xu, X. W., et al. (2018). ECG signal classification method based on deep learning and fuzzy C-means. Acta Automatica Sinica, 44(10), 1913–1920.

[9] Tao, L., Liu, B. N., & Liang, W. (2021). Automatic arrhythmia detection based on CNN-LSTM hybrid model. Journal of Shandong University (Engineering Science), 51(3), 30–36.

[10] Ye, S. C., Xu, C. H., & Qiao, Q. L. (2020). Research on arrhythmia classification based on residual network fusion model. Journal of Shaanxi Normal University (Natural Science Edition), 48(6), 10–17.

[11] Jin, Y. R., Qin, C. J., Liu, J. L., et al. (2021). A novel incremental and interactive method for actual heartbeat classification with limited additional labeled samples. IEEE Transactions on Instrumentation and Measurement, 70, 2507212-1–2507212-12.

[12] Qin, L., Xie, Y. T., Liu, X. C., et al. (2021). An end-to-end 12-leading electrocardiogram diagnosis system based on deformable convolutional neural network with good antinoise ability. IEEE Transactions on Instrumentation and Measurement, 70, 2508313-1–2508313-13.

[13] Liu, Z., Wang, H., Gao, Y., et al. (2021). Automatic attention learning using neural architecture search for detection of cardiac abnormality in 12-lead ECG. IEEE Transactions on Instrumentation and Measurement, 70, 2513612-1–2513612-12.

[14] Jin, Y., Li, Z., Qin, C., et al. (2023). A novel attentional deep neural network-based assessment method for ECG quality. Biomedical Signal Processing and Control, 79, 269–280.

[15] Roopa, C. K., & Harish, B. S. (2020). Automated ECG analysis for localizing thrombus in culprit artery using rule based information fuzzy network. International Journal of Interactive Multimedia and Artificial Intelligence, 6(1), 16–25.