ELECTROCARDIOGRAM ABNORMAL DETECTION MODEL USING MACHINE LEARNING APPROACH

Today, datasets can be obtained transparently and freely. It can also be used to categorize and predict diseases with high-risk factors. Moreover, the extracted datasets can generate important information for the entire population if handled accurately. This dataset can predict heart disease using a machine learning approach with explicit calculations. We compared the prediction of abnormal electrocardiograms in this study using machine learning with three algorithms, namely support vector machine (SVM), k-Nearest Neighbors (KNN), and multilayer perceptron (MLP) classifier. We used 14 attributes: (1) age, (2) systolic, (3) heart rate, (4) obesity, (5) smoking, (6) alcohol, (7) exercise, (8) treadmill exercise results, (9) total cholesterol, (10) high-density lipoprotein, (11) low-density lipoprotein, (12) creatinine, (13) serum glutamic oxaloacetic transaminase, and (14) urine protein. The results predict the indicated heart disease and display the accuracy of each algorithm. Furthermore, the results revealed that the machine learning technique employing the KNN algorithm is the most effective, with an accuracy rate of 89.375%.