Heart disease prediction using techniques of classification in machine learning

Abstract

In this thesis we have examined the accuracy of various classifiers to predict heart disease and heart vessel blockage. We have also analyzed the key features contribut ing to heart vessel blockage. We have used a dataset containing 14 attributes related to heart disease of 1025 patients. From our study we found that the Decision Tree, Random Forest and KNN algorithm gave the highest accuracy for detecting heart disease. For predicting heart vessel blockage, the Decision tree had the highest accuracy. While analyzing the features contributing to heart vessel blockage, we found that patients’ age and cholesterol level has the highest contribution. Hence, monitoring the patient’s cholesterol level may help prevent heart vessel blockage.