Advanced Heart Disease Prediction Using Fuzzy-Rough Sets and Enhanced Missing Data Imputation Techniques

For early diagnosis and efficient clinical decision-making, heart disease prediction accuracy is essential. However, the reliability of predictive models is severely hampered by the prevalence of missing data in medical datasets. The fuzzy-rough set theory-based improved missing data imputation system presented in this work is intended to address the ambiguity and incompleteness of medical data. Our approach builds upon the Cardiovascular Disease Multiple Imputation Technique (CVDMIT) by combining fuzzy-rough sets with sophisticated classifiers such as Random Forest and new ensemble learning methods. Several benchmark datasets, including the UCI Heart Disease dataset, were used to validate the method, which showed a 95% accuracy rate higher than more conventional techniques like expectation maximization and fuzzy C-means. Our proposed method achieves better performance through Extensive experiments which enhance sensitivity while improving precision and recall metrics. Our research establishes fundamental principles for integrating fuzzy-rough set-based imputation into clinical workflows thus enabling better scalability of heart disease prediction models.