Integrating Symbolic Regression and Photoplethysmography for Monitoring Blood Pressure Estimation
Resumo
This paper advances non-invasive blood pressure (BP) monitoring by leveraging photoplethysmography signals, enhanced through the integration of symbolic regression (SR) and traditional machine learning techniques. Our novel methodology combines traditional SR-based and feature extraction methods, utilizing recursive feature elimination with cross-validation (RFECV) for optimal feature selection. Comparative analysis across extensive datasets shows that integrating SR with RFECV enhances model transparency and predictive accuracy, providing clinically interpretable mathematical expressions that improve our understanding of BP estimation dynamics, which is crucial for healthcare diagnostics.
Referências
Bittner, V. (2020). The new 2019 aha/acc guideline on the primary prevention of cardiovascular disease. Circulation, 142(25):2402–2404.
Burlacu, B. (2023). Gecco’2022 symbolic regression competition: Post-analysis of the operon framework. In Proceedings of the Companion Conference on Genetic and Evolutionary Computation, pages 2412–2419.
Burlacu, B., Kronberger, G., and Kommenda, M. (2020). Operon c++: An efficient genetic programming framework for symbolic regression. In Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion, GECCO ’20, page 1562–1570, New York, NY, USA. Association for Computing Machinery.
Cava, W. G. L., Orzechowski, P., Burlacu, B., de França, F. O., Virgolin, M., Jin, Y., Kommenda, M., and Moore, J. H. (2021). Contemporary symbolic regression methods and their relative performance. CoRR, abs/2107.14351.
de França, F. O. (2023). Alleviating overfitting in transformation-interaction-rational symbolic regression with multi-objective optimization. Genetic Programming and Evolvable Machines, 24(2):13.
González, S., Hsieh, W.-T., and Chen, T. P.-C. (2023). A benchmark for machine-learning based non-invasive blood pressure estimation using photoplethysmogram. Scientific Data, 10(1):149.
Guyon, I. and Elisseeff, A. (2003). An introduction to variable and feature selection. Journal of machine learning research, 3(Mar):1157–1182.
Kachuee, M., Kiani, M. M., Mohammadzade, H., and Shabany, M. (2015). Cuff-less high-accuracy calibration-free blood pressure estimation using pulse transit time. In 2015 IEEE int. symp. on circuits and systems (ISCAS), pages 1006–1009. IEEE.
Kearney, P. M., Whelton, M., Reynolds, K., Muntner, P., Whelton, P. K., and He, J. (2005). Global burden of hypertension: analysis of worldwide data. The lancet, 365(9455):217–223.
La Cava, W. G., Lee, P. C., Ajmal, I., Ding, X., Solanki, P., Cohen, J. B., Moore, J. H., and Herman, D. S. (2023). A flexible symbolic regression method for constructing interpretable clinical prediction models. NPJ Digital Medicine, 6(1):107.
Liang, Y., Chen, Z., Ward, R., and Elgendi, M. (2018). Photoplethysmography and deep learning: enhancing hypertension risk stratification. Biosensors, 8(4):101.
Makke, N. and Chawla, S. (2024). Interpretable scientific discovery with symbolic regression: a review. Artificial Intelligence Review, 57(1):2.
Rudin, C. (2019). Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nature machine intelligence, 1(5):206–215.
Schmidt, M. and Lipson, H. (2009). Distilling free-form natural laws from experimental data. science, 324(5923):81–85.
Seidyo Imai Aldeia, G. and Olivetti de Franca, F. (2024). Interpretability in symbolic regression: a benchmark of explanatory methods using the feynman data set. arXiv e-prints, pages arXiv–2404.
Wong, M. K. F., Hei, H., Lim, S. Z., and Ng, E. Y. K. (2023). Applied machine learning for blood pressure estimation using a small, real-world electrocardiogram and photo-plethysmogram dataset. Mathematical Biosciences and Engineering.