Heart Disease Analysis and Prediction Using EDA and ML Classifiers

Heart is one of the most important organs as blood is pumped throughout the body by the heart. Heart disease is widely regarded as the illness that kills individuals most quickly. To prevent catastrophic risks and diminish heart-related problems, prompt identification of heart disease is essential. The disease prediction system helps clinicians identify serious illnesses as early as possible. This research paper aims to provide a complete examination of the features contributing to heart disease with respect to by means of exploratory data analysis and predict the possibility of heart disease with the use of machine learning algorithms. Through EDA, it was discovered that, of all these features, the number of major vessels (0–3) stained by fluoroscopy, different types of chest pain, greatest heart rate reached, exercise-induced angina, exercise-induced ST depression compared to rest, and inclination of the peak exercise ST segment are the ones that have the greatest impact on the diagnosis of heart disease. Utilising the UCI heart disease dataset containing the features, three machine learning classifiers are implemented: decision tree (DT), random forest (RF), and K-nearest neighbour (KNN). When compared to other heart disease prediction models, Random Forest has the best rate of accuracy (85.25%).