A hybrid approach to heart disease prediction using a fractional-order mathematical model and machine learning algorithm.

Abstract

Heart disease remains one of the leading causes of morbidity and mortality worldwide, necessitating the development of more accurate and efficient diagnostic tools. This study presents a hybrid approach to heart disease prediction, combining fractional-order dynamics with decision tree algorithms and an interactive graphical user interface (GUI). Fractional-order models allow for a more elaborate representation of the complex physiological processes involved in heart disease, including factors such as cholesterol levels, blood pressure, inflammation, and plaque buildup. By integrating decision trees, a machine learning (ML) method known for its interpretability and efficiency in classification tasks, this approach enhances predictive accuracy. The use of an interactive GUI further enables healthcare professionals to visualize and interact with the model, providing real-time insights into patient risk profiles. The model's fractional-order differential equations (FDEs) account for varying rates of progression in different health parameters, offering a dynamic view of heart disease risk. Comprehensive simulations demonstrate the efficacy of the model, which outperforms traditional prediction models in terms of both accuracy and usability. The hybrid framework is intended to serve as a robust tool for clinicians, offering an innovative combination of advanced mathematical modeling and user-friendly machine-learning techniques for heart disease prediction. Our findings show that the decision tree classifier performed well, with 93% accuracy, 95% precision, 90% recall, and an F1-score of 0.92. The model handled non-linear relationships and missing data effectively, achieving an ROC-AUC score of 0.99. Key correlations, such as between ST depression and exercise-induced angina, were identified. Fractional-order simulations revealed how cholesterol, blood pressure, and other factors influenced heart disease risk, reinforcing clinical links through numerical simulations.