LOCAL ADAPTATION OF MRA-FW FOR CARDIOVASCULAR RISK ASSESSMENT IN ARAL SEA REGION INPATIENTS
DOI:
https://doi.org/10.5281/zenodo.23364166Abstract
This study examines the adaptation of the Medical Risk-Aware Feature Weighting (MRA-FW) algorithm to the clinical profile of cardiology inpatients from the Aral Sea region. A total of 2,039 inpatient medical records from Nukus (2024–2026) were retrospectively digitized and analyzed. High proportions of cardiovascular, cerebrovascular, and renal comorbid conditions were observed; these values describe the selected inpatient cohort and are not interpreted as population prevalence. The original MRA-FW formulation combines model-based feature importance, statistical relevance, and a medical prior. For regional adaptation, a regularized combination of local and global medical priors is proposed to incorporate regional clinical information without excessively shifting the prior coefficients. In the previous reproducible MRA-FW experiment using 68,597 cleaned Cardiovascular Disease records and five classifiers, Recall ranged from 0.9432 to 0.9803 and F2-score from 0.8215 to 0.8401. The principal effect of the method is interpreted as a sensitivity-oriented shift of the screening operating point rather than a universal improvement in all metrics.Keywords
Aral Sea region, cardiovascular risk, inpatient cohort, comorbidity, artificial intelligence, MRA-FW, adaptive feature weighting, medical prior, Recall, F2-score, CardioAI-DSSReferences
1. Collins, G. S., Moons, K. G. M., Dhiman, P., et al. (2024). TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378.
2. Liu, T., et al. (2025). Machine learning based prediction models for cardiovascular disease risk using electronic health records data: systematic review and meta-analysis. European Heart Journal - Digital Health, 6(1), 7–22.
3. Uzokov, J., Alyavi, A., Alyavi, B., & Abdullaev, A. (2024). How artificial intelligence can assist with ischaemic heart disease. European Heart Journal, 45(21), 1866–1868. https://doi.org/10.1093/eurheartj/ehae030
4. Azimova, N. D., Ashirbaev, S. P., & Vikhrov, I. P. (2020). The first steps in artificial intelligence development in medicine in Uzbekistan. Health Problems of Civilization, 14(4), 314–319. https://doi.org/10.5114/hpc.2020.98086
5. Saito, T., & Rehmsmeier, M. (2015). The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE. 2015;10(3):e0118432.
6. Ulianova, S. (2026). Cardiovascular Disease Dataset. Kaggle. 70,000 patient records, 11 features and binary target. Accessed 2026.
7. Pedregosa, F., Varoquaux, G., Gramfort, A., et al. (2011). Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12, 2825–2830.
8. Ronco, C., Haapio, M., House, A. A., et al. (2008). Cardiorenal syndrome. Journal of the American College of Cardiology, 52(19), 1527–1539.
9. Dehkordi, P., Tavakolian, K., Rocco, T., Fozilov, K., Yuldashov, A., Yuldashov, B., Atamuratov, B., et al. (2026). Electro-mechanical risk score for coronary artery disease risk estimation using seismocardiography. European Heart Journal - Digital Health. 2026;7(8):ztag139. DOI: 10.1093/ehjdh/ztag139. ztag139. DOI: 10.1093/ehjdh/ztag139. https://doi.org/10.1093/ehjdh/ztag139
10. Chen, T., & Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference. 2016:785–794.
11. Lundberg, S. M., & Lee, S. I. (2017). A Unified Approach to Interpreting Model Predictions. Advances in Neural Information Processing Systems, 30, 4765–4774.
Downloads
Published
How to Cite
Issue
Section
