Performance Analysis of Heart Disease Detection Using Different Machine Learning Approaches

The leading cause of illness and mortality globally, heart disease is a serious global health problem. For successful prevention and intervention, timely and precise cardiac disease prediction is essential. Using the well-known Cleveland dataset from the UCI repository, we underwent an extensive performance examination of several ML approaches in the context of heart disease prediction in this study. The 303 cases and 14 features in this collection include essential clinical factors. The KNN, Random Forest, Naive Bayes, AdaBoost, Decision Tree, and Logistic Regression machine learning algorithms are among those being examined. Each is being assessed for its predictive efficacy. Our work reveals fascinating insights into algorithm performance through substantial data preparation, robust model training, and a thorough assessment utilizing measures like accuracy, sensitivity, specificity, and PPV, NPV. Notably, our results show that the KNN algorithm is the most effective at predicting heart disease by 93%. This finding has implications for improving clinical decision support systems and early heart disease detection, thereby lowering the prevalence of the condition and its negative effects on public health. This research highlights the crucial contribution of ML to the development of cardiovascular medicine and the significance of algorithm choice for predicting accuracy.