Comparative analysis of Machine Learning Techniques for the Detection of Heart Disease

The World Health Organization stated that an alarming 32 percent of worldwide fatalities was from heart related challenges. This percentage is even more escalated in poor and moderate countries. Most heart related issues can be ameliorated or managed if they are detected on time. Hence, with the advent of machine learning techniques, prediction of heart related issues can be made. Several machine learning methods have been proposed for heart related issues prediction, and each with varying performance level. A few comparisons have been made amongst these machine learning techniques with little details. In this study, a detailed comparison of four machine learning techniques was considered. The machine learning methods considered were selected based on their performance from literatures. The examined methods include the decision tree, random forest, support vector machines, and logistic regression. Prior to training and testing, the dataset was preprocessed by encoding the categorical data and scaling the dataset features. The outcome of the testing showed that the accuracy was obtained for logistic regression and support vector machine; with both having an accuracy of 0.882.