Heart Disease Risk Prediction Using Machine Learning and Explainable Ai : A Data-Driven Approach for Early Diagnosis and Prevention

Accurate and interpretable heart disease prediction is of vital importance for preventive diagnosis and cost-effective healthcare decision-making. The work presented here focuses on an Enhanced SSL framework based on the SimCLR contrastive paradigm, integrating deep SE-ResNet encoding, multihead attention , and Focal Loss optimization to offer robust and interpretable cardiovascular risk prediction. Using the Heart Statlog Cleveland–Hungary dataset , comprising 1,190 clinical records and 12 diagnostic attributes, the framework learned discriminative representations through contrastive pretraining followed by supervised fine tuning. A comprehensive business analytics component was incorporated for evaluating clinical and economic impacts . Risk based screening and cost–benefit analysis revealed that targeted high-risk screening strategies significantly outperformed universal screening in both detection efficiency and financial return, thus supporting evidence based resource allocation. Market opportunity estimation further highlighted large-scale potential for AI-driven preventive cardiology. Explainability through SHAP was performed , which identified medically consistent determinants such as ST slope, chest pain type, and exercise angina as the most influential features. Accuracy was 94.97%, the F1-score was 95.29%, and ROC-AUC was 0.9729 for the proposed model, outperforming traditional Random Forest and SVM baselines. Results hereby validate that self-supervised representation learning, coupled with explainable AI and business analytics, enhances predictive precision, interpretability, and real-world healthcare utility.