Sisir Kumar Rajbongshi

DOI: https://doi.org/10.5281/zenodo.24131866

Feature relevance for stroke risk is usually judged with crisp statistics or post-hoc model explanations, neither of which describes a risk factor in the graded, linguistic terms clinicians use. This study ranks the candidate features of a public, de-identified stroke dataset (5,109 patients, 4.87% with stroke) with a fuzzy soft set discrimination score (FSDS), which sums the absolute differences between the mean Low, Medium and High memberships of stroke and non-stroke patients. The score is shown to be twice the total variation distance between the two classes’ term distributions, which puts continuous and binary features on one scale; for a binary feature it is twice the difference in prevalence. Because crude scores mix risk factors with proxies for age, an age-adjusted score is introduced, computed within age strata and tested by permutation within strata. Without adjustment, “ever married” ranked second and “work type: children” sixth; after adjustment both lost significance (q = 0.16 and 0.84), and only average glucose level, hypertension, BMI and heart disease remained significant beside age (q ≤ 0.048). In ten repetitions of five-fold cross-validation, with ranking repeated inside every training fold, the top eight FSDS features matched the full feature set and ANOVA, chi-square and mutual-information selection (logistic regression ROC-AUC 0.840 against 0.838–0.841; corrected resampled t-test, all p > 0.4). With class weighting, balanced random forests or SMOTE, recall was 42%–80%. FSDS selections were less stable than ANOVA selections in small samples. The contribution is interpretability and confound control, not higher accuracy.

Fuzzy soft sets, feature selection, stroke risk, confounding, class imbalance, interpretable machine learning

A fuzzy soft set discrimination score for ranking stroke risk features was examined on a severely imbalanced public dataset. The score equals twice the total variation distance between the stroke and non-stroke distributions over linguistic terms, which justifies ranking continuous and binary features together and gives each score a readable profile. Without adjustment the score, like the ANOVA, chi-square and mutual-information filters, ranked age proxies such as marital status among the top features; with age stratification and a stratified permutation test it recovered exactly the four established modifiable risk factors (glucose, hypertension, heart disease and BMI) beside age. In a leakage-free evaluation its selections predicted stroke as well as all 17 features and as well as the standard filters, and they improved the SVM and KNN substantially, but they were not more accurate than the filters and were less stable in small samples. Future work should test the adjusted score on independent cohorts with more continuous clinical measurements and adjust for several confounders at once.

[1] GBD 2019 Stroke Collaborators. (2021). Global, regional, and national burden of stroke and its risk factors, 1990–2019: A systematic analysis for the Global Burden of Disease Study 2019. The Lancet Neurology, 20(10), 795–820. https://doi.org/10.1016/S1474-4422(21)00252-0
[2] Feigin, V. L., Brainin, M., Norrving, B., Martins, S., Sacco, R. L., Hacke, W., et al. (2022). World Stroke Organization (WSO): Global stroke fact sheet 2022. International Journal of Stroke, 17(1), 18–29. https://doi.org/10.1177/17474930211065917
[3] Dubey, Y., Tarte, Y., Talatule, N., Damahe, K., Palsodkar, P., & Fulzele, P. (2024). Explainable and interpretable model for the early detection of brain stroke using optimized boosting algorithms. Diagnostics, 14(22), 2514. https://doi.org/10.3390/diagnostics14222514
[4] Dritsas, E., & Trigka, M. (2022). Stroke risk prediction with machine learning techniques. Sensors, 22(13), 4670. https://doi.org/10.3390/s22134670
[5] Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems 30 (pp. 4765–4774).
[6] Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). “Why should I trust you?”: Explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1135–1144). https://doi.org/10.1145/2939672.2939778
[7] Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8(3), 338–353. https://doi.org/10.1016/S0019-9958(65)90241-X
[8] Molodtsov, D. (1999). Soft set theory—First results. Computers & Mathematics with Applications, 37(4–5), 19–31. https://doi.org/10.1016/S0898-1221(99)00056-5
[9] Maji, P. K., Biswas, R., & Roy, A. R. (2001). Fuzzy soft sets. Journal of Fuzzy Mathematics, 9(3), 589–602.
[10] Sanchez, E. (1979). Inverse of fuzzy relations. Application to possibility distributions and medical diagnosis. Fuzzy Sets and Systems, 2(1), 75–86. https://doi.org/10.1016/0165-0114(79)90017-4
[11] Çelik, Y., & Yamak, S. (2013). Fuzzy soft set theory applied to medical diagnosis using fuzzy arithmetic operations. Journal of Inequalities and Applications, 2013, 82. https://doi.org/10.1186/1029-242X-2013-82
[12] Singh, R., & Bhardwaj, N. (2025). Fuzzy soft set theory applications in medical diagnosis: A comprehensive review and the roadmap for future studies. New Mathematics and Natural Computation, 21(2), 597–619. https://doi.org/10.1142/S1793005725500279
[13] Jensen, R., & Shen, Q. (2004). Semantics-preserving dimensionality reduction: Rough and fuzzy-rough-based approaches. IEEE Transactions on Knowledge and Data Engineering, 16(12), 1457–1471. https://doi.org/10.1109/TKDE.2004.96
[14] Zhang, X., Mei, C., Li, J., Yang, Y., & Qian, T. (2023). Instance and feature selection using fuzzy rough sets: A bi-selection approach for data reduction. IEEE Transactions on Fuzzy Systems, 31(6), 1981–1994.
[15] Ben Kahla, M., Kanzari, D., Ben Amor, S., Ayachi Ghannouchi, S., & Martinho, R. (2025). Enhanced fuzzy score-based decision support system for early stroke prediction. ACM Transactions on Computing for Healthcare, 6(1), 1–23. https://doi.org/10.1145/3703461
[16] fedesoriano. (2021). Stroke prediction dataset [Data set]. Kaggle. https://www.kaggle.com/datasets/fedesoriano/stroke-prediction-dataset
[17] Ambroise, C., & McLachlan, G. J. (2002). Selection bias in gene extraction on the basis of microarray gene-expression data. Proceedings of the National Academy of Sciences, 99(10), 6562–6566. https://doi.org/10.1073/pnas.102102699
[18] Kaufman, S., Rosset, S., Perlich, C., & Stitelman, O. (2012). Leakage in data mining: Formulation, detection, and avoidance. ACM Transactions on Knowledge Discovery from Data, 6(4), 15. https://doi.org/10.1145/2382577.2382579
[19] Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16, 321–357. https://doi.org/10.1613/jair.953
[20] Guyon, I., & Elisseeff, A. (2003). An introduction to variable and feature selection. Journal of Machine Learning Research, 3, 1157–1182.
[21] Chandrashekar, G., & Sahin, F. (2014). A survey on feature selection methods. Computers & Electrical Engineering, 40(1), 16–28. https://doi.org/10.1016/j.compeleceng.2013.11.024
[22] Kraskov, A., Stögbauer, H., & Grassberger, P. (2004). Estimating mutual information. Physical Review E, 69(6), 066138. https://doi.org/10.1103/PhysRevE.69.066138
[23] Kuncheva, L. I. (2007). A stability index for feature selection. In Proceedings of the 25th IASTED International Multi-Conference: Artificial Intelligence and Applications (pp. 390–395).
[24] Nogueira, S., Sechidis, K., & Brown, G. (2018). On the stability of feature selection algorithms. Journal of Machine Learning Research, 18(174), 1–54.
[25] Rajbongshi, S. K., & Talukdar, D. (2013). An introduction to a family of fuzzy subsets forming Boolean algebra. International Journal of Computer Applications, 68(24).
[26] Rajbongshi, S. K., & Talukdar, D. (2013). Some aspects of fuzzy Boolean algebra formed by fuzzy subsets. International Journal of Advanced Research in Computer Science and Software Engineering, 3(7).
[27] Rajbongshi, S. K., & Gogoi, P. (2023). On some theoretic aspects of fuzzy subsets. International Journal for Research in Applied Science & Engineering Technology, 11(2).
[28] World Health Organization. (2000). Obesity: Preventing and managing the global epidemic. Report of a WHO consultation (WHO Technical Report Series 894). World Health Organization.
[29] American Diabetes Association Professional Practice Committee. (2024). 2. Diagnosis and classification of diabetes: Standards of care in diabetes—2024. Diabetes Care, 47(Suppl. 1), S20–S42. https://doi.org/10.2337/dc24-S002
[30] Mantel, N., & Haenszel, W. (1959). Statistical aspects of the analysis of data from retrospective studies of disease. Journal of the National Cancer Institute, 22(4), 719–748.
[31] Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B, 57(1), 289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x
[32] He, H., & Garcia, E. A. (2009). Learning from imbalanced data. IEEE Transactions on Knowledge and Data Engineering, 21(9), 1263–1284. https://doi.org/10.1109/TKDE.2008.239
[33] Chen, C., Liaw, A., & Breiman, L. (2004). Using random forest to learn imbalanced data (Technical Report 666). Department of Statistics, University of California, Berkeley.
[34] Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., et al. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12, 2825–2830.
[35] Lemaître, G., Nogueira, F., & Aridas, C. K. (2017). Imbalanced-learn: A Python toolbox to tackle the curse of imbalanced datasets in machine learning. Journal of Machine Learning Research, 18(17), 1–5.
[36] 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, 10(3), e0118432. https://doi.org/10.1371/journal.pone.0118432
[37] Nadeau, C., & Bengio, Y. (2003). Inference for the generalization error. Machine Learning, 52(3), 239–281. https://doi.org/10.1023/A:1024068626366
[38] Bouckaert, R. R., & Frank, E. (2004). Evaluating the replicability of significance tests for comparing learning algorithms. In Advances in Knowledge Discovery and Data Mining (PAKDD 2004), Lecture Notes in Computer Science, 3056 (pp. 3–12). Springer. https://doi.org/10.1007/978-3-540-24775-3_3

Sisir Kumar Rajbongshi (2026). Fuzzy Soft Sets for Stroke Risk Feature Selection: Age-Adjusted Ranking, Clinical Validity and Robustness to Severe Class Imbalance. International Journal of Computer Techniques, 13(6), 721–734. ISSN: 2394-2231. DOI: https://doi.org/10.5281/zenodo.24131866