Predicting the Occurrence of Major Adverse Cardiac Events in Patients with Acute Coronary Syndrome Using Synthetic Minority Oversampling Technique and Machine Learning Approach

Abstract

Cardiovascular disease (CVD) is the leading cause of mortality globally. Therefore, the early prediction and prevention of the CVD are very important. The objective of this paper is to predict the occurrence of Major Adverse Cardiac Events (MACE) in CVD patients using machine learning approach based on the Korea Acute Myocardial Infarction Registry (KAMIR-NIH) dataset. After extracting the target dataset, the occurrence number of patients with MACE accounted for only 5.5% of the total. Hence, in the experiment, we first used the Synthetic Minority Oversampling Technique (SMOTE) to overcome the obvious imbalanced problem, then we generated four popularly used machine learning-based prediction models: Random forest, logistic regression, extreme gradient boosting (XGBoost), support vector machine, and try to discover the best prediction model. The overall performance of those prediction models was evaluated based on the accuracy, precision, recall, F1-score, and the area under the curve (AUC). As a result, we found the random forest-based prediction model can achieve the highest accuracy (AUC = 0.9601) for the prediction of MACE occurrence.