Fuzzy Mutual Information Based Hybrid Classification Algorithm for Enhanced Heart Attack Prediction

The early detection of heart attacks necessitates the use of precise and dependable models, which are vital to the provision of cardiovascular health care. In this study, we provide a unique hybrid classification algorithm based on fuzzy mutual information that is optimized for better heart attack prediction. To create a baseline for comparison, conventional models such as Naive Bayes, Support Vector Machine, K-Nearest Neighbor, Random Forest, Extreme Gradient Boost, and Logistic Regression were used. For every conventional model, the assessment indicators were measured, especially accuracy. However, our hybrid approach showed notable increases in efficiency by integrating mutual information-based fuzzy feature selection. In particular, the suggested method outperformed the accuracy scores of the conventional models, achieving a remarkable 93.257%. The proposed algorithm's capacity to identify pertinent features is enhanced by the addition of fuzzy mutual information-based feature selection, yielding a more accurate and perceptive predictive model for cardiovascular risk assessment.