KAN-Type Neural Network Model for ECG Signal Analysis and Classification
DOI:
https://doi.org/10.15276/opu.2.72.2025.17Keywords:
electrocardiogram, multilayer perceptron, MLP-KAN, additional layer, wavelet transformation, radial basis functions, classification, accuracyAbstract
Electrocardiogram analysis is critically important for the diagnosis of cardiovascular pathologies, as precise identification of local signal features ensures reliable arrhythmia classification. The relevance of this work is determined by the need to improve the accuracy of automated systems capable of processing signal segments of varying length under noisy conditions. The objective of the study is the development of a model based on a multilayer perceptron with an additional layer that combines radial basis functions and wavelet transformation for accurate detection of local signal characteristics. The tasks included preparation and normalization of signal segments, experimental comparison of different types of wavelets and their decomposition levels, selection of the number of functions in the additional layer, and assessment of the impact of architectural parameters on classification accuracy and computational performance. The research methods involved processing signals from an open database, normalizing segment values within a defined range, constructing models using a multiclass loss function and gradient descent optimization, and evaluating performance through classification accuracy and balanced accuracy metrics for segments with ventricular rhythm disturbances. The results demonstrated that integrating an additional layer with radial basis functions and wavelet transformation increases arrhythmia classification accuracy, ensures model stability under variations in segmentation parameters and noise presence, and allows effective extraction of local signal features. The use of different wavelet types and decomposition levels enables achieving an optimal balance between accuracy and computational efficiency. The scientific novelty lies in combining radial basis functions with wavelet transformation to enhance the detection of local electrocardiogram signal features, providing increased noise robustness. The practical significance of the study is the creation of an effective methodology for automated electrocardiogram analysis with high arrhythmia classification accuracy.
Downloads
References
Hemanth, D. J., Gupta, D., & Balas, V. E. (2019). Intelligent data analysis for biomedical applications. Academic Press. DOI: https://doi.org/10.1016/C2017-0-03676-5.
Xiao, Q., Lee, K., Mokhtar, S. A., Ismail, I., Md Pauzi, A. L., Zhang, Q., & Lim, P. Y. (2023). Deep learning-based ECG arrhythmia classification: A systematic review. Applied Sciences, 13(8), Article 4964. DOI: https://doi.org/10.3390/app13084964.
Khalid, M., Pluempitiwiriyawej, C., Wangsiripitak, S., & Abdulkadhem, A. A. (2024). The applications of deep learning in ECG classification for disease diagnosis: A systematic review and meta-data analysis. Engineering Journal, 28(8), 45–77. DOI: https://doi.org/10.4186/ej.2024.28.8.45.
Galchonkov, O., Baranov, O., Maslov, O., Babych, M., & Baskov, I. (2025). MLP-K AN: Implementation of the Kolmogorov-Arnold layer in a multilayer perceptron. Eastern-European Journal of Enterprise Technologies, 3/4(135), 34–41. DOI: https://doi.org/10.15587/1729-4061.2025.328928.
Ahmed, A. F., & Al-Obaidi, M. K. (2022). A review of ECG signal filtering approaches. Global Journal of Engineering and Technology Advances, 11(3), 093–097. DOI: https://doi.org/10.30574/gjeta.2022.11.3.0099.
Sharma, N., & Sidhu, J. S. (2016). Removal of noise from ECG signal using adaptive filtering. Indian Journal of Science and Technology, 9(48). DOI: https://doi.org/10.17485/ijst/2016/v9i48/106424.
Shcherbakova, G., & Koshutina, D. (2025). Development of a criterion for selecting the level of wavelet decomposition for QRS detection in electrocardiogram signals using energy and entropy. Proceedings of Odessa Polytechnic University, 1(71), 157–166. DOI: https://doi.org/10.15276/opu.1.71.2025.18.
Pan, J., & Tompkins, W. J. (1985). A real-time QRS detection algorithm. IEEE Transactions on Biomedical Engineering, BME-32(3), 230–236. DOI: 10.1109/TBME.1985.325532.
PhysioNet. (n.d.). MIT-BIH Arrhythmia Database. Retrieved from https://www.physionet.org/content/mitdb/1.0.0.
Sörnmo, L., & Laguna, P. (2005). Electrocardiogram signal processing. Wiley.
He, H., Tan, Y., & Wang, Y. (2015). Optimal base wavelet selection for ECG noise reduction using a comprehensive entropy criterion. Entropy, 17(9), 6093–6109. DOI: https://doi.org/10.3390/e17096093.
Agrawal, P., Arun, V., & Basu, A. (2025). Artificial neural network based ECG feature extraction using wavelet transform. In Emerging Wireless Technologies and Sciences. ICEWTS 2024 (CCIS, Vol. 2399, pp. 8–22). Springer, Cham. DOI: https://doi.org/10.1007/978-3-031-87886-2_2.
Kaverinskiy, V., Chaikovsky, I., Mnevets, A., & Malakhov, K. S. (2025). Scalable clustering of complex ECG health data: Big data clustering analysis with UMAP and HDBSCAN. Computation, 13(6), Article 144. DOI: https://doi.org/10.3390/computation13060144.
Shcherbakova, G., Antoshchuk, S., Koshutina, D., & Sakhno, K. (2024). Adaptive clustering for distribution parameter estimation in technical diagnostics. 12th International Conference on Applied Innovations in IT (ICAIIT), Köthen, Germany. DOI: https://doi.org/10.25673/115650.
Lourenço, A., Carreiras, C., Bulò, S. R., & Fred, A. (2014, August 25–29). ECG analysis using consensus clustering [Paper presentation]. 2014 22nd European Signal Processing Conference (EUSIPCO), Lisbon, Portugal.
Iacobescu, P., Marina, V., Anghel, C., & Anghele, A. (2024). Evaluating binary classifiers for cardiovascular disease prediction: Enhancing early diagnostic capabilities. Journal of Cardiovascular Development and Disease (JCDD), 11(12), Article 396. DOI: https://doi.org/10.3390/jcdd11120396.
Guo, C., Ahmed, S., & Alouini, M.-S. (2023). Machine learning-based automatic cardiovascular disease diagnosis using two ECG leads. arXiv. DOI: https://doi.org/10.48550/arXiv.2305.16055.
Wu, Z., & Guo, C. (2025). Deep learning and electrocardiography: Systematic review of current techniques in cardiovascular disease diagnosis and management. BioMed Eng OnLine, 24, Article 23. DOI: https://doi.org/10.1186/s12938-025-01349-w.
Savalia, S., & Emamian, V. (2018). Cardiac arrhythmia classification by multi-layer perceptron and convolution neural networks. Bioengineering, 5(2), Article 35. DOI: https://doi.org/10.3390/bioengineering5020035.
Li, Z. (2024). Kolmogorov-Arnold networks are radial basis function networks. arXiv. Retrieved from https://arxiv.org/abs/2405.06721.
Liu, J., Xia, X., Han, C., Hui, J., & Feng, J. (2022). Classification of arrhythmia based on convolutional neural networks and encoder-decoder model. Computer Systems Science and Engineering. DOI: https://doi.org/10.32604/cmc.2022.029227.
Mavaddati, S. (2024). ECG arrhythmias classification based on deep learning methods and transfer learning technique. Biomedical Signal Processing and Control. DOI: https://doi.org/10.1016/j.bspc.2024.107236.
Wang, M., Feng, Y. T., Guan, S., & Qu, T. (2024). Multi-layer perceptron-based data-driven multiscale modelling of granular materials with a novel Frobenius norm-based internal variable. Journal of Rock Mechanics and Geotechnical Engineering. DOI: https://doi.org/10.1016/j.jrmge.2024.02.003.
Hadj Kilani, B. (2024). Kolmogorov-Arnold networks: Key developments and uses. Qeios. DOI: https://doi.org/10.32388/7NNCAA.
Yu, R., Yu, W., & Wang, X. (2024). KAN or MLP: A fairer comparison. arXiv. DOI: https://doi.org/10.48550/arXiv.2407.16674.
Liu, Z., Wang, Y., Vaidya, S., Ruehle, F., Halverson, J., Soljačić, M., Hou, T. Y., & Tegmark, M. (2024). KAN: Kolmogorov–Arnold Networks. arXiv. Retrieved from https://arxiv.org/abs/2404.19756.
Ismayilova, A., & Ismayilov, M. (2023). On the universal approximation property of radial basis function neural networks. arXiv. Retrieved from https://arxiv.org/abs/2304.02220.
Ta, T. H. (2024). BSRBF-KAN: A combination of B-splines and radial basis functions in Kolmogorov-Arnold networks. ResearchGate. DOI: https://doi.org/10.13140/RG.2.2.16755.54562.
Hollósi, J. (2024). Efficiency analysis of Kolmogorov–Arnold networks for visual data processing. SMTS 2024 Conference Proceedings. DOI: https://doi.org/10.3390/engproc2024079068.
Seghouane, A.-K., & Shokouhi, N. (2021). Adaptive learning for robust radial basis function networks. IEEE Transactions on Cybernetics, 51(5), 2847–2856. DOI: https://doi.org/10.1109/TCYB.2019.2951811.
Panda, S., & Panda, G. (2022). On the development and performance evaluation of improved radial basis function neural networks. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 52(6), 3873–3884. DOI: https://doi.org/10.1109/TSMC.2021.3076747.
Bozorgasl, Z., & Chen, H. (2024). Wav-KAN: Wavelet Kolmogorov-Arnold networks. arXiv. Retrieved from https://arxiv.org/abs/2405.12832.
Koshutina, D. V., & Shcherbakova, H. Y. (2025, May 15–16). Automated selection of wavelet decomposition level based on energy and entropy for QRS complex detection in ECG [Paper presentation]. 15th International Scientific Conference for Students and Young Scientists "Modern Information Technologies – 2025", Odesa, Ukraine. DOI: https://doi.org/10.5281/zenodo.15521809.
