Heart Stroke Prediction Using Machine Learning Approach with Clinical and Demographic Factors

Heart strokes remain a critical global health issue, with early detection being the key to improving outcomes. Increasing stroke rates and the need for effective treatments pose major challenges to sustainable healthcare. Machine learning plays a pivotal role in predicting optimal cardiovascular treatment strategies. Using a dataset of 1,320 patient records—including heart rate, blood pressure, blood sugar, Troponin, CK-MB, and other medical parameters—three models (Random Tree, AdaBoost, and Decision Stump) were evaluated. AdaBoost achieved the highest accuracy (98.47%), followed by the Random Tree (96.88%) and Decision Stump (87.35%). The superior performance of AdaBoost highlights its effectiveness in capturing complex clinical data patterns. Although Random Tree and Decision Stump offer some utility, they fall short in modelling intricate relationships. This study underscores the potential of machine learning to intensifies cardiovascular care by enabling personalized, data-driven treatment planning and improving both efficiency and patient outcomes.