Predicting Death from Heart Disease: A Comparative Analysis of Machine Learning Algorithms

The circulatory system's maintenance is a complex and essential function performed by the human heart. However, heart disease has been the leading cause of global mortality for many years, and it comprises various disorders that significantly impact this vital organ. Therefore, early detection of heart disease is crucial, and it requires a reliable, precise, and efficient method. In this regard, researchers and medical practitioners have employed several machine learning algorithms, including Decision Tree, Random Forest, Gradient Boosting, Logistic Regression, Support Vector Machine, and Naive Bayes, to analyze a heart disease dataset from the UCI Machine Learning Repository. Hyperparameter tuning was also incorporated into the individual algorithms to improve model performance. The study aimed to assess the outcomes of various machine learning algorithms for predicting death from heart disease. Upon conducting a 10-fold cross-validation, the Random Forest algorithm performed the best, with an accuracy rate of 98%. This outcome demonstrates the potential of machine learning in developing accurate and reliable models to detect heart diseases at an early stage, ultimately saving lives. The results of this study are significant in terms of early detection of heart diseases and their prevention. The use of machine learning algorithms has shown promising results, and further development can improve the accuracy of the models. This information can be helpful to medical practitioners, policymakers, and the general public and can help save numerous lives globally.