Empowering heart attack treatment for women through machine learning and optimization techniques

Heart attack detection and treatment in women remain significantly under-optimized due to differences in symptom presentation and physiological characteristics compared to men, leading to delayed or incorrect diagnoses. Addressing this gap, this study introduces an optimized ensemble learning approach that leverages a novel voting classifier combining the Waterwheel Plant Algorithm (WWPA) with Stochastic Fractal Search (SFS). The proposed WWPA+SFS model is designed to enhance the accuracy of heart attack classification in women by integrating multiple machine learning classifiers, including Gaussian Naive Bayes, Random Forest, Logistic Regression, Stochastic Gradient Descent Classifier, Support Vector Classifier, Decision Tree, and k-nearest Neighbors. A comprehensive clinical dataset comprising 303 patient records and 14 features—covering demographic data, exercise-induced angina, chest pain type, major vessel count, cholesterol levels, fasting blood sugar, and resting electrocardiographic results—was used for evaluation. The model’s performance was validated using 10-fold cross-validation, Analysis of Variance (ANOVA), and the Wilcoxon Signed Rank Test, benchmarking it against other optimization-based classifiers such as Grey Wolf Optimization (GWO), Whale Optimization Algorithm (WOA), Particle Swarm Optimization (PSO), and Genetic Algorithm (GA). The proposed WWPA+SFS model achieved the highest classification accuracy (97.01%) and demonstrated low variance across multiple trials. These results underline the robustness and effectiveness of the proposed method in optimizing diagnostic models for women’s cardiovascular care, potentially reducing misdiagnosis rates, lowering healthcare costs, and contributing to personalized treatment advancements in clinical practice.