Heart Disease Prediction Using Machine Learning and Data Mining Techniques: Application of Framingham Dataset

Abstract

Introduction:Heart problems have gained a lot of interest in medical research because of their impact on human health where early diagnosis is critical to delaying the development of heart disease, the world's leading cause of death. thus, it is much needed to predict the possibility of occurrence of heart disease based on their attributes. Objective:This research aims into a variety of machine learning classification algorithms for predicting heart disease. Methods:The 10-fold cross-validation resampling is used to validate the prediction model. Aim and the prediction scores of each algorithm are evaluated with performance metrics such as prediction accuracy, confusion matrix, F1-Meuser, and
suggested geometric mean. Results:It was revealed that classifying the HD dataset using different classification algorithms produces extremely promising results, with a classification accuracy of 83.95, 84.5, 84.82, 84.89, and 85.05 % for the KNN, SVM, DT, LR, and RF algorithms, respectively. The RF algorithm successfully predicts 85.05 % (true positive rate) of the deceased cases correctly.
Conclusion:This study suggests that the RF method predicts Framingham possibilities better than other algorithms for the smaller (4240 records) dataset, based on the findings of other machine learning classification techniques on the Framingham dataset.