“…According to the Framingham heart study [ 7 ], the prediction models for prediction of heart diseases have mainly been categorized into two groups, such as regression-based risk assessment methods and machine learning-based classification methods. The previous risk assessment methods have been the Framingham Risk Score (FRS) [ 8 , 9 , 10 ], QRISK [ 11 , 12 ], Thrombolysis in Myocardial Infarction (TIMI) [ 13 , 14 ], Global Registry of Acute Coronary Events (GRACE) [ 15 , 16 ], and History, Electrocardiogram, Age, and Risk factors and Troponin (HEART) [ 17 , 18 , 19 ] models, whereas the machine learning-based classification methods for heart disease [ 20 ] are Random Forest (RF) [ 21 , 22 , 23 ], Extra Tree (ET) [ 20 , 24 ], Support Vector Machine (SVM) [ 25 ], Gradient Boosting Machine (GBM) [ 26 , 27 ], Neural Networks (NN) [ 28 , 29 , 30 ], and other ensemble models [ 6 , 31 ]. However, there are some drawbacks and limitations of previous models, as follows.…”