Heart disease detection using ensemble and non-ensemble machine learning methods

Cardiovascular diseases are one of the leading causes of disability and death. In 2019, heart disease caused the death of approximately 17.9 million people worldwide, representing 32% of all deaths recorded worldwide. Machine learning has emerged as one of the most well-known areas in computer science. Machine learning has been addressing many complex problems, especially in the medical field, with remarkable success. This study aims to detect heart diseases using ensemble and non-ensemble machine learning models and feature selection methods. A dataset titled "Heart Disease Dataset" obtained from IEEE DataPort was used in this study. The dataset was analyzed and preprocessed, and then the most relevant features were selected using three combined feature selection methods. Various non-ensemble machine learning methods such as KNN, random forest, XGB and GBM, and ensemble machine learning methods such as voting and stacking were applied. According to the results, the random forest model achieved the best score with 92.4% accuracy.