Classifying Patients with Myocardial Infarction and Heart Failure by Using SVM and KNN Learning Techniques

Abstract

Cardiovascular diseases (CVD) are considered to be the leading cause of death globally and millions of people from all around the world die annually due to the different types of heart diseases. There are multiple major and minor risk factors that together contribute to developing heart disease. These risk factors include age, sex, tobacco, physical inactivity, genetics etc. Therefore, it’s hard to predict heart disease in patients using conventional methods. On the other hand however, with the help of technology, it has now become easier to achieve this goal. The process begins by evaluating datasets containing patient’s risk factors. Then, the evaluated datasets would be analyzed using one of the many machine learning techniques. Finally, the analyzed data would be used as a base for classifying and predicting heart disease in new patients. In this paper, we used two of the most advanced machine learning techniques Support Vector Machine (SVM) technique as well as K-Nearest Neighbor (KNN) to analyze the data that we obtained from 210 patients in Sulaimani Cardiac Hospital between (October 16^th,2019 to January 9^th, 2020). In conclusion, we obtained that the SVM yields more accurate results (82.6%) compared to the KNN method (73.0%).